Fiktív hősei a honfoglalás során pedig kitalált csatákat vívtak elképzelt népek és a honfoglalás korában a Kárpát-medencében nem létező hatalmak ellen.

“Its fictional heroes fought fictional battles against imagined peoples and powers that did not exist in the Carpathian Basin at the time of the conquest.”

(Wikipedia, Gesta Hungarorum)

What stands out in the highlighted Hungarian words is simply their featuring dental phonemes near the end of the word, and their congruence with the expression of the past: as the English translation testifies, these are the verbs conjugated in the past tense and past participles in the short passage (with the exception of the verb to be, from which an irregularity can be anticipated).

Let's understand more closely the Hungarian sentence, as always by identifying with English word group by word group. First, we highlight the minimal sentence, stripped of all modifiers, adverbial adjuncts of time and the long adverbial complements starting with against.

Fiktív hősei a honfoglalás során pedig kitalált csatákat vívtak elképzelt népek és a honfoglalás korában a Kárpát-medencében nem létező hatalmak ellen.

This kernel translates into …his heroes fought battles… with hősei (his heroes) a nominative possessive form of hős (3rd person singular possessor with plural possessed), vívtak (fought) our verb in the past indicative and csatákat (battles), as might be expected, an accusative plural. honfoglalás further refers to the Hungarian Conquest and our translation omitted the redundancy highlighted below.

Fiktív hősei a honfoglalás során pedig kitalált csatákat vívtak elképzelt népek és a honfoglalás korában a Kárpát-medencében nem létező hatalmak ellen.

során, an adverb, means during and kor-á-ban the inessive possessive of kor (time), so, in its time. (A very literal (improper) translation of a honfoglalás korában would be *in the Hungarian Conquest its time.)

Fiktív hősei a honfoglalás során pedig kitalált csatákat vívtak elképzelt népek és a honfoglalás korában a Kárpát-medencében nem létező hatalmak ellen.

Here, we have …fought battles (csatákat vívtak) against (ellen) people and powers (népek és hatalmak)… where the preposition ellen calls the nominative (plural) in népek and hatalmak.

Fiktív hősei a honfoglalás során pedig kitalált csatákat vívtak elképzelt népek és a honfoglalás korában a Kárpát-medencében nem létező hatalmak ellen.

Share

The present participle létező is an attributive modifier of hatalmak and the clause it leads translates in to powers that did not exist in the Carpathian Basin (literally, *in the Carpathian Basin not existing powers) - which leave us with our past and past participles of interest:

Fiktív hősei a honfoglalás során pedig kitalált csatákat vívtak elképzelt népek és a honfoglalás korában a Kárpát-medencében nem létező hatalmak ellen.

kitalált is the past participle of the verb kitalál (to figure out, to invent, to make up) and is here attribute of csatákat, so invented (fictive) battles. Let's feel its conjugation table. The present tense (jelen idő) indefinite makes

én kitalálok
te kitalálsz
ő kitalál
mi kitalálunk
ti kitaláltok
ők kitalálnak

and we do notice the addition to the stem of a dental sound in the past tense (múlt idő) indefinite

én kitaláltam
te kitaláltál
ő kitalált
mi kitaláltunk
ti kitaláltatok
ők kitaláltak

Fiktív hősei a honfoglalás során pedig kitalált csatákat vívtak elképzelt népek és a honfoglalás korában a Kárpát-medencében nem létező hatalmak ellen.

The same can be said of elképzelt, past participle of elképzel (to imagine) whose past indefinite makes elképzelt in the third person singular. Eventually, vívtak is the third person plural past indefinite of vív. Its present counterpart is vívnak and the past “dental-suffix” is here clear too.

Subscribe now

This already well-established trend is confirmed by the following extracts from the chronicles of Hungarian history.

Anonymus volt az első, aki az Attilától való származás lehetőségét megfogalmazta.

“Anonymus was the first who formulated the possibility of descent from Attila.”

(Wikipedia, Gesta Hungarorum)

Ennek a királynak az ivadékából sarjadt az igen nevezetes és roppant hatalmú Attila király.

“From this king's lineage descended the very famous and immensely powerful King Attila.”

(Gesta Hungarorum, ca. 1200)

Akkor a választásuk arra esett, hogy majd Pannónia földjét keresik fel. Erről ugyanis a szállongó hírből azt hallották, hogy az Attila király földje, akinek az ivadékából Álmos vezér, Árpád apja származott.

“Then their choice fell on seeking the land of Pannonia. For they had heard from rumors that it was the land of King Attila, from whose line descended Chief Álmos, Árpád's father.”

(Gesta Hungarorum, ca. 1200)

Az Úr megtestesülése utáni négyszázegyedik, a magyarok Pannóniába történt bejövetelétől számított huszonnyolcadik esztendőben a magyarok, vagyis a hunok a rómaiak szokása szerint egyetértő akarattal királyul emelték maguk fölé Attilát, Bendegúz fiát, aki előbb a kapitányok közé tartozott;

“In the four hundred and first year after the Lord's incarnation, and in the twenty-eighth year after the Hungarians' entry into Pannonia, the Hungarians, or rather the Huns, following Roman custom, by unanimous will raised Attila, son of Bendegúz, who had previously belonged among the captains, as king above themselves.”

(Képes Krónika, 1358)

The Képes Krónika (Chronicon Pictum), from which the last passage is taken, continues to recount the Hungarian myth of origins. The collection of illustrated historical chronicles written by Marc de Kalt in 1358, on commission from King Louis I, is based on a lost manuscript of the Gesta Ungarorum dating from the end of the 11th century and the time of Saint Ladislas. It endeavors to recount the continuity of the Hun and Hungarian reigns. Unequivocally, emelték and tartozott are the two past forms translated by raised and had belonged respectively. By contrast, the use of the past participles történt and számított, not explicitly rendered in the English translation, requires clarification. történt (happened) is used as an attributive adjective to the ablative bejövetelétől (from the entry): történt bejövetelétől is literally (and improperly) *from the happened entry [into Pannonia], that is, more regularly, after the entry [into Pannonia] had happened, where a natural and fluid translations further omits had happened, as it is redundant with the preposition after. Similarly számított (counted) is employed as an attributive adjective to esztendőben. Whereas the suffix -től betrayed the ablative, -ben signals the inessive, the case of the space in which we are. huszonnyolcadik esztendőben means in the 28th year and a fluent translation here again omits the excess zeal in a literal *in the 28th counted year.

Ultimately, the past tense, the grammar book tell us, “is expressed with the suffix -t or -ott/-ett/-ött and inflects for person and number.”

Share Old Norse Linguist

We have spoken at length about the dental suffix that runs through the formation of past tense in Germanic languages, a legacy of Proto-Germanic. Observing the phenomenon outside the spectrum is at least intriguing, and calls for further comparative studies.

Here again, we prefer historical accounts, for their several virtues. Firstly, they abound in past-tense forms, which are presently our main interest. Second, they are particularly propitious for comparative linguistics: they teem with proper nouns and dates, highly recognizable from one language to another, which act as landmarks, secure rivets on which parallel exegesis can be fixed. The following concerns the peace negotiations between the Ottoman Empire and the Kingdom of Hungary in 1444:

6 Mart 1444'te ilk müzakereler başladı. Ardından 24 Nisan 1444'te Kral Ladislas II. Murad'a müzakereleri kabul ettiğini belirten bir mektupla birlikte elçisi Stojka Gisdaniç'i Edirne'ye gönderdi. Esirlerin teatisinde varılan uzlaşının ardından toprak meseleleri müzakere edildi. Osmanlılar, Macarların 1443-1444 seferinde kaybettikleri Güvercinlik (Golubaç) ve Semendire'nin iadesini temine çalıştılar. Ancak Karaman Beyi İbrahim'in Anadolu'daki saldırıları nedeniyle apar topar 12 Haziran'da mevcut koşullarda barışa razı oldular.

Perusing the text, however sybaritic before translation, is enough to intrigue us. The final words of the five sentences are not entirely random: they have similarities of form. başladı, gönderdi, edildi all end in -di while çalıştılar and oldular share the ending -lar, or even -d/t + vowel + lar. A sentence-ending verb cannot come as a surprise to a Germanist, of whom the learner of Old Norse is a close relative. If these five Turkish forms were our past tense verbs, we would certainly be holding our dental suffixes. Let's take a look at the English translation, and indulge in our favorite exercise: parallel reading and comprehension. We highlight verbs in the past tense.

“The initial negotiations began on March 6, 1444. Then on April 24, 1444, King Ladislas sent his envoy Stojka Gisdanić to Edirne with a letter indicating to Murad II that he accepted the negotiations. After reaching an agreement on the exchange of prisoners, land issues were negotiated. The Ottomans tried to secure the return of Güvercinlik (Golubac) and Smederevo, which the Hungarians had lost in their 1443-1444 campaign. However, due to the attacks of Ibrahim, Bey of Karaman, in Anatolia, they hastily agreed to peace under the existing conditions on June 12.”

The first sentence is clear: under our hypothesis, başladı is the verb began. That leaves a date, 6 Mart 1444'te, and a subject noun phrase, ilk müzakereler for the initial negotiations. Let's test our proficiency in a language we didn't know a few hours ago on the second one. We emphasize everything we recognize transparently.

Ardından 24 Nisan 1444'te Kral Ladislas II. Murad'a müzakereleri kabul ettiğini belirten bir mektupla birlikte elçisi Stojka Gisdaniç'i Edirne'ye gönderdi.

“Then on April 24, 1444, King Ladislas sent his envoy Stojka Gisdanić to Edirne with a letter indicating to Murad II that he accepted the negotiations.”

We also assumed that gönderdi to be a past tense verb. Its position near Stojka Gisdaniç'i, which would then serve as its direct object, speaks for its being the main clause’s verb, sent. Ardından could well start the time adjunct given its situation (then). This leaves his envoy […] with a letter indicating […] that he accepted the negotiations to our mapping speculation. elçisi is a good candidate for [Stojka Gisdaniç’s] envoy given the proximity in location and ending, -i. The treasure hunt continues. If elçisi Stojka Gisdaniç'i is indeed his envoy, object of sent, the -i ending might be typical of the direct object case. The exegesis of our first sentence had expedited the correspondence of ilk müzakereler with the initial negotiations. müzakereleri is an obvious match for negotiations, this time in the accusative as a direct object of accepted. All that remains is with a letter indicating […] that he accepted. Firstly, we can assume with a letter to be closest to the phrase its accompanies, elçisi Stojka Gisdaniç'i (his envoy Stojka Gisdanić) and that he accepted closest to its presumed direct object, müzakereleri (the negociations). Secondly, the ending of ettiğini could again betray an accusative. In with a letter indicating […] that he accepted, a letter is no accusative, maybe an instrumental case. The only element that can carry the accusative is acceptance, a nominalized flavor of the verb phrase that he accepted, direct object to indicating. From there on, we are content with assessing electronically our guesses. bir mektupla birlikte is indeed with a letter, -la marking the instrumental case. In kabul ettiğini, kabul is the substantive acceptance, the prefix et- verbalizes it while -tiğ- forms the past participle and -ini marks the accusative possessive. Literally, something such as *his (having) accepted. The suffix -en in belirten marks the present participle, indicating.

Incidentally, we increased our lexicon, inferred possible desinences of the nominative (plural), accusative and instrumental, observed how we verbalize a noun, and form the possessive and the past participle, among others. It is in this way, guided by curiosity and bilingual reading, that we believe it is most delectable and effective to learn a new language - rather than through primary, rebarbative exercises ordered in booklets for novices. We have set some foundations. We may forget Turkish for a long time, but if we come back to it, we will already feel a little at home.

We can now draw some conclusions from our observations. The past tense of Turkish verbs (at least those we have encountered) is systematically formed with stem + suffix di/du + person marker (possibly null). We have namely

• başla-dı-ø: stem (başla-) + past suffix (-dı) + null person marker

• gönder-di-ø: stem (gönder-) + past suffix (-di) + null person marker

• edil-di-ø: stem (edil-) + past suffix (-di) + null person marker

• çalış-tı-lar: stem (çalış-) + past suffix (-tı) + person marker (-lar)

• ol-du-lar: stem (ol-) + past suffix (-du) + person marker (-lar)

where the person markers -ø and -lar mark the 3rd person singular and plural respectively. (An attentive mind will wonder about başla-dı-ø's lack of congruence with its plural subject ilk müzakereler. Turkish, it seems, takes some license with subject-verb agreement, and the singular is preferred for an inanimate plural subject. Here, the subject can also be interpreted as a collective singular concept: English can after all also say the negotiations began.)

Ultimately, the past tense suffix in Turkish will be chosen among

-di/-dı/-du/-dü
-ti/-tı/-tu/-tü

based on some vowel harmony rules. A watered-down form of vowel harmony is well known to practitioners of Old Norse or Icelandic: the phonetic u- and i-omlyd (umlaut), which we shall review shortly. In short, a word's vowels (especially during declension) adjust in height, frontness or roundedness, to be more “in tune” with each other.

Share Old Norse Linguist

The Hungarian-Ottoman peace of 1444 is soon breached by Hungarian King Władysław III. The battle of Varna proves fatal to him: Murad II’s triumph paves the way for the Ottoman takeover of Constantinople in 1453. Ottomans too are mindful of dynastic historiography. While Hungarians trace a Hunnic lineage through the Gesta Hungarorum, Murad II models himself after the legendary Ghazi kings:

He drew from the noble behavior of the nameless Caliphs in the Battalname, an epic about a fictional Arab warrior who fought against the Byzantines, and modelled his actions on theirs. He was careful to embody the simplicity, piety, and noble sense of justice that was part of the ghazi king persona. […]

Murad II successfully painted himself as a simple soldier who did not partake in royal excesses, and as a noble ghazi sultan who sought to consolidate Muslim power against non-Muslims such as the Venetians and Hungarians.

History books and myths also served imperial Japan. The Kojiki (古事記, Archives of Ancient Affairs) and the Nihon Shoki (日本書紀, Chronicles of Japan) pass for the oldest chronicles: the former is commissioned by the imperial reign and completed in the early 8th century. It legitimizes Yamato's rule by a divine lineage, and noble families of a new rank society by a chosen genealogy.

Let’s read an account for an episode concerning Emperor Jimmu, the legendary first emperor of Japan according to the Kojiki.

そして龍田へ進軍するが道が険阻で先へ進めなかった。そこで東へ軍を向けて胆駒山を経て中洲（大和国）へ入ろうとし、この地を支配する長髄彦と孔舎衛坂で戦った。 戦いに利なく、長兄の五瀬命は流れ矢に当たって負傷した。そして日の神の子孫の自分達が日に向かって（東に向かって）戦うことは天の意思に逆らうことだと悟ることとなった。彦火火出見尊は兵を集めて草香津まで退き、再び海路南へと向かった。

Guided by our Germanist thoughts, we notice a strange pattern: (almost) all sentences finish with a similar syllable,った (-tta). A closer look further shows a certain recurrent vicinity of the symbols か and な,

そして龍田へ進軍するが道が険阻で先へ進めなかった。そこで東へ軍を向けて胆駒山を経て中洲（大和国）へ入ろうとし、この地を支配する長髄彦と孔舎衛坂で戦った。 戦いに利なく、長兄の五瀬命は流れ矢に当たって負傷した。そして日の神の子孫の自分達が日に向かって（東に向かって）戦うことは天の意思に逆らうことだと悟ることとなった。彦火火出見尊は兵を集めて草香津まで退き、再び海路南へと向かった。

Here is an electronic translation:

“When they advanced to Tatsuta, the road was too treacherous to proceed. They then turned the army east, passing through Mount Ikoma, attempting to enter Nakasu (Yamato Province), and fought with Nagasunehiko who controlled this area at Kosae slope. The battle was not favorable, and their eldest brother Gose no Mikoto was wounded by a stray arrow. They then realized that fighting while facing the sun (facing east) as descendants of the sun deity was against heaven's will. Hikohohodemi no Mikoto gathered his troops, retreated to Kusaka-tsu, and again headed south by sea.”

Let's decipher the sibylline

そして龍田へ進軍するが道が険阻で先へ進めなかった。

by means of our ordinary approach to language learning. The piece translates into

When they advanced to Tatsuta, the road was too treacherous to proceed.

Asking semiconductors for a translation of “they advanced to Tatsuta” we are given 龍田へ進軍する, which fits our passage perfectly.

そして龍田へ進軍するが道が険阻で先へ進めなかった。

We don’t find here the recurrentった that we might hope to be a past tense marker. The verb 進軍する , we learn, is indeed a narrative present tense. Next, a proposed translation for “the road was treacherous” is 道が険阻だった which we spot almost entirely in the original

そして龍田へ進軍するが道が険阻で先へ進めなかった。

道が険阻 may be assume to match “the road… treacherous” while the verb asks for clarification. In your remnant 進めなかった, な (na) marks the negation, 進め is the modal can advance, and the whole なかった (nakatta) marks the negative past tense. In fact, the Japanese seems to literally say *the road treacherous […] could not advance where something indicating causality is clearly missing: the role is filled by the conjunctionで (de) which can be translated by being:

…the road being treacherous, (they) could not advance

Incidentally, we have illustrated once more how a parallel inspection, driven by curiosity, of new language and English makes the opaque clearer. We claim this the most fruitful and funniest way to learn new languages generally.

Subscribe now

Upset by Germanist ideas, we have been struck so far by a strange coincidence: that of dental suffixes for past tense formation in Hungarian, Turkish, and Japanese. As mentioned earlier, a dental suffix runs through the Germanic spectrum in the formation of the past tense of weak verbs, and the pattern is assumed, by Ringe (2017) in particular, to be a Proto-Germanic legacy. (Here, we were looking at how the Proto-Germanic weak verb classes identified by Ringe prolong into Old Norse.) Proto-Germanic is assumed to have formed its past indicative and subjunctive with a *-d- (perhaps a *-dēd-) suffix. Its past indicative would have followed the following pattern:

Past tense indicative suffix-ending(s)

Sg -d-ǭ
-d-ēz
-d-ē

Pl -d-ū
-d-u-diz
-d-u-m, -d-u-d, -d-u-n

Share Old Norse Linguist

We shall soon come back to this Proto-Germanic phenomenon and the efforts that have been made to shed light on its origins. But for now, let's focus on our coincidence. Apart from a certain dental suffix in the formation of the past tense, Hungarian, Turkish and Japanese have in common that they share no Indo-European genealogy. Consensus places Hungarian in the Ugric branch of a Uralic family and derives it from a Proto-Ugric spoken “from the end of the 3rd millennium BC until the first half of the 1st millennium BC, in Western Siberia, east of the southern Ural Mountains”. Turkic languages on the other hand are thought to have derived from some Proto-Turkic that would have emerged between the Transcaspian steppe and Manchuria around 2500 BC. The so-called Japonic languages encompass Japanese and the indigenous languages of the Ryukyu Islands, that are assumed to have originated from a Proto-Japonic brought to the archipelago in the early 4th century BC by settlers stemming from the Korean peninsula.

Nevertheless, the history of linguistics abounds in hypotheses and controversies around a potential kinship between these three language families, with a recent resurgence of interest fueled by the unprecedented computational possibilities of modern times. By the late seventeenth century already, the outlines of an Altaic family were being discerned, grouping together Turkic, Mongolic and Tungusic. The 18th century proposed that it be rallied by Uralic languages on the one hand, and Japanese and Korean on the other. Let's first consider Japanese language specialist Alexander Vovin's arguments against the latter hypothesis, as he shifted from advocating to criticizing it. In short, he denounces the lack of sufficient evidence, and charges vehemently (Vovin, 2009) certain recent scholarly works whose attempts at proof he finds fundamentally flawed. Here we are at the very core of the dispute: what proof is necessary and sufficient to demonstrate an indisputable kinship between two (families of) languages? According to Vovin (2011, p.6; emphasis added):

… the demonstration of a [genealogical] relationship is only possible by either showing the existence of common paradigmatic morphology, and/or by demonstrating the existence of regular phonetic correspondences within the basic vocabulary.

Regular morphological correspondences between two words (from different languages) being compared are not enough : such correspondences must also not be due to borrowing; be “predictive-productive”, i.e. be systematic and extend to inflection and overall word formation based on active (productive) rules (thereby predictive of new word formation); leave no unaccounted segment; come with semantic correspondence, supported not by bare word-lists but philological evidence, corpus occurrences, and cultural context. As such restrictions demand hard work, it is easy to be lulled by similarities that are mere borrowings or happenstance; and hasty judgment could justify nearly any adoption by the Altaic family. If we were to follow Robbeets' reasoning in its ultimate consequences, Vovin scorns, we would add Russian: Robbeets claims a Proto-Altaic denominal verbal (i.e. formative of verbs from substantives) suffix *-la- on the basis of a Proto-Japonic *-ra- yielding Old-Japonic -r-, Turkic *-lA-, Mongolic *-lA-, and Tungusic *-lA-. Vovin further notes the similarity between Russian kamla- “to shamanize” and Turkic qam-la- “to shamanize” derived from qam “shaman”. Those verbs form their respective past as follows

and Vovin (2011, p.1-2) concludes

… stems are identical but the paradigms are not: derivational verbal morphology is easily borrowed, but inflectional is not.

In short, a conclusive resemblance must extend to inflection. The borrowing (rather than the inheritance) of a derivation rule is further betrayed by its loose systematism. Mongolian has boγorla- “cut the throat” visibly corresponding to Old Turkic boγuz-la- with same meaning derived from boγaz~ boγuz “throat”; Mongolian ögütle- “advise” to Old Turkic ögüt-le- “advise” from ögüt “advice”; or Mongolian čiŋla- ~ čiŋna- “listen” to Old Turkic tïŋ-la- “listen” from tïŋ “listening”. Yet, Vovin tells us, there are no Mongolic root words meaning throat, advice, or listening which boγorla-, ögütle- or čiŋla- could be derived from by appending a suffix -la: these verbs were borrowed, and only later did the rule become productive in Mongolian, with, for example, üge-le- “to say” from üge “word”, as the -la/-le suffix was eventually “segmented and began to be used after pure Mongolian nominal roots”.

Establishing a language family is an ambitious task, as it faces nearly impossible criteria. For Giorgio Orlandi (2020, p.43), another detractor of an Altaic Japanese, “one of the universal criteria for demonstrating the integrity of a language family”, captured in one formula, is “confidentially eliminating chance, borrowing, and other non-genetic factors”. We near tautology: to demonstrate that a resemblance attests to kinship, one must eliminate the possibility of all alternative explanatory factors, all of which are, by definition, non-genetic. Later, Orlandi (2020, p.52) specifies: two languages are genetically related if the ancestor language, the protolanguage from which both derive, is “recoverable only by means of specific linguistic techniques which seek to reconstruct [the] rules and restructurings" that mold their respective evolution after the branching off. These techniques are mainly internal and external reconstruction. Internal reconstruction compares, internally to one language, “allomorphs in order to arrive at a reconstructed invariant form of a morpheme” while external reconstruction, the “comparative method”, seeks to reconstruct the ancestor from word matching across languages. Published in 2003, the Etymological Dictionary of the Altaic Languages (Statosin et all.), “a huge work in three volumes, containing 2,800 etymologies” is thus regarded by its authors as the definitive end of the debate:“the very fact that it is possible to compile a dictionary of common Altaic heritage, they say, appears to be a proof of the validity of the Altaic theory”. Yet Orlandi warns: reconstruction can easily lapse into “teleological exercise” or a “speculative etymological construct”. It is particularly Statosin's phonetic reconstruction that Orlandi seem to dislike. The sound laws governing the derivation of the daughter languages must be both typologically plausible and consistent with the principles of naturalness - established pattern of language evolution are observed - and minimality - each sound change affects only one phonetic feature at a time. And Orlandi concludes: Statosin's reconstruction is not sufficient, but neither is reconstruction necessary.

While the Trans-Himalayan family is widely accepted (Greenberg 1996: 134, LaPolla 2001: 225), a complete reconstruction of Proto-Trans-Himalayan is still unavailable at the moment. This proves that the “reconstruction” of a Proto-language does not imply the validity of that language family, and vice versa.

It seems that Altaists weave by day the demonstration that anti-Altaists unweave by night: like Penelope's knitwear, the scholarly talk can go on for a long time.

Subscribe now

We understand that an unsettling similarity between two languages of humanity can be due to heritage, borrowing or chance - and that proof can be epic. But are we sure that nothing is missing from the picture? Some interlanguage similarities come to our minds, which none of the three above-mentioned factors primarily explains: pathways of grammaticalization. Gradually, usage transmutes, via abstraction, a lexical element into a grammatical one, a content-word into a function-word. Thus, numerous markers of future tense across languages derive, along recurrent morphological, semantic and phonetic paths, from verbs of movement or intention.

Heine and Kuteva (2002) catalogue a host of such semantic source-target associations: too much coincidence to invoke luck, while anachronistic universality defies both borrowing from contact and heredity. As we have seen, grammaticalization arises from communication needs. The paths of species evolution require the increasingly effective expression of increasingly abstract notions, which find their way by metaphor from the concrete: the time of movement projects into the future, so that progressive usage fixes the English “is going to”, the Dutch “gaan”, the Swedish “komma att” as future tense auxiliaries. The constants of the grammaticalization phenomenon and its modalities - source-target pairs, morphological, phonetic and semantic paths - thus respond to species modalities of cognition and communication, and to their evolutionary trajectories sculpted by interdependent biology and culture. Among these cognitive universals, we also find the hypothetical “universal grammar”, an innate propensity for language and an innate notion of its functional categories. According to a Strong Minimal Hypothesis, what makes human language unique is the hierarchical structure of its syntax: the emergence of its merge function, together with the cognitive and sensory-motor capacities underpinning thought and expression, propel our species on its distinctive trajectory. (It makes it possible to assemble the and apple in the set {the, apples}, which further combines to ate in {ate,{the, apples}}, etc.) How Could Language Have Evolved? (Bolhuis, Tattersall, Chomsky & Berwick, 2014), where interests from neurobiology, anthropology and linguistics converge, seem to call on future science to clarify the potential genomics of such merge function.

With merge […] the basic properties of human language emerge. Evolutionary analysis can thus be focused on this quite narrowly defined phenotypic property, merge itself, as the chief bridge between the ancestral and modern states for language. Since this change is relatively minor, it accords with what we know about the apparent rapidity of language’s emergence.

Although its malfunctions “produce speech deficits in modern people”, the gene FOXP2, the authors recall or deplore, can no longer be “regarded as “the” gene “for” language” as many others intermingle in “its normal expression”. The desire is palpable to find its theoretical replacement, a gene of the decisive merge function. Eventually, beyond the brain, another obvious biological universal is that of the vocal tract, which conditions the phonological possibilities and drifts of languages.

Share Old Norse Linguist

In short, it is not surprising that languages resemble each other: they are human. Perhaps there is a biology of past tense formation, a genetics of the past tense forming dental suffixes that at least four very different kind of languages seem to feature. Let's sketch out two ideas for pursuing our investigation. We have taken a close look at the grammaticalization of future markers: our suffix could be a result of that of past expression. One of the constant processes of grammaticalization, as we have seen, is semantic, morphosyntactic, and phonetic coalescence, which the African dialects of !Xun illustrate well at work: the Proto-!Xun (reconstructed) verbs of motion *gǀè (come) and *ú (go) merge with the conjunctions *kà and *tà or the transitive suffix *-ā to produce future tense markers, for instance in ò-tāǁé (will die) and oga gǀyee (will come) that respectively derive from *ú tā ǁé (go and die) and *ú kā gǀè (go and come).

Let's imagine our dental past tense markers originating from the same slow coalescence and abstraction from some kind of lexical items with concrete meaning. Their phonetic similarity across languages could then result from that of their source words. What could justify interlanguage congruence of sounds and meaning? Well, here also, theories and literature abound. Modern “sound symbolism” and its sophisticated laboratory experiments continue to enliven the didactic disputes of Plato's Cratylus.

References

Geréb, L. (Trans.). (1993). Képes Krónika: A magyarok régi és legújabb tetteiről, eredetükről és növekedésükről, diadalaikról és bátorságukról. (T. Tarján, Ed.). Magyar Hírlap és Maecenas Kiadó.

Ringe, D. (2017). From Proto-Indo-European to Proto-Germanic (2nd ed.). Oxford University Press.

Vovin, A. (2009). Japanese, Korean, and Other 'Non-Altaic' Languages. Central Asiatic Journal, 53(1), 105-147.

Vovin, A. (2011). Why Japonic is not demonstrably related to 'Altaic' or Korean. Paper presented at the 20th International Conference on Historical Linguistics (ICHL XX), Osaka, Japan, July 30, 2011.

Heine, B., & Kuteva, T. (2002). World Lexicon of Grammaticalization. Cambridge University Press.

Bolhuis, J. J., Tattersall, I., Chomsky, N., & Berwick, R. C. (2014). How Could Language Have Evolved? PLOS Biology, 12(8), e100193434.

Theorists call grammaticalization the linguistic process that gradually metamorphoses lexical items into grammatical ones. We have recently, by way of illustration, visited modal verbs from Old to Modern English, and evoked the frequent intermingling of the expression of tense and modality. The future tense auxiliary will, in particular, bears also modal nuances. It has been grammaticalized as a future tense marker by analogy with its predecessor's lexical meaning - intention and willingness.

Another typical pathway in the grammaticalization of the future tense originates from verbs of movement. This is what we shall be looking at more closely - in the present essay, we will evoke the phenomenon across several languages, especially of the Germanic spectrum, and this will lead us in an upcoming part three to a study of the future tense in Old Norse.

Subscribe now

Reconstructing a Proto-!Xun, a hypothetical ancestor of the !Xun language spoken by hunter-gatherers in south-west Africa, no trace of “a conventionalized future tense form or construction” can be found, Heine and his co-authors tell us. (Heine, Kuteva & Narrog, 2017, p.4-9) In contrast, Proto-!Xun seems to have featured the verbs of motion *ú (go) and *gǀè (come), whose descendants participate to the formation of the future tense in today's !Xun dialects. Most probably, Heine concludes, these verbs preexisted the grammatical markers of the future tense in the history of the !Xun, and the latter derived from the former: “movement-based future tenses” developed. The following is a repertoire of !Xun dialects and their future tense markers.

W1 and E3 clearly developed their markers from the Proto-!Xun *gǀè (come). N1, N2, C1, E2, perhaps W2 and W3, from *ú (go). We also find the conjunctions tā and oga, which will soon prove pivotal in the discussion. In addition, the !Xun dialects can draw on three types of constructions to form the future tense. These differ in the type of association between the two verbs - auxiliary and lexical - that they involve. A closer look at these three types of construction reveals much about the ongoing process of grammaticalization, from verbs of movement to future tense markers. We find indeed

1. A complement-based future where the suffix -á serves to augment the valency of the verb, that is, its aptitude to take on a complement, e.g. turning an intransitive verb into a transitive one. Here is an example from the N1 dialect, with Heine’s encoding. (T refers to preposition/suffix, FUT to future marker, TOP to topic marker)

2. A serializing future that juxtaposes both verbs, here in the E2 dialect

3. A particle-based future where future markers bear resemblance with but differ from gǀè / ú. Namely we can find oga, ò-tā, óá, forms that are undergoing phonetic, syntactic, morphological coalescence. In the following W2 example

the future marker óá results (most probably, Heine says) from *ú-ā > *ó-á > óá where -á is the above (1) complement-friendly suffix. Similarly, ò-tā and oga can certainly be reconstructed back to a coalescence of *ú with the conjunctions tà and kà respectively.

Here comes the most interesting part. Our suffix -á and conjunctions tà/kà are not limited to forming the future tense. -á adapts, in general, a verb to take on a complement. This can be a verb in the past tense (past tense marked by kè), as in (W2)

hȁ má kè ú- á ḿ. (He went to eat.)

Likewise, tà and kà serve coordination in general, as in (W2)

hȁ má kè ú kā m̋. (He went and ate.)

Consequently, we can assume future tense constructions to still be semantically colored by their lexical counterparts, coexisting in the dialects. For instance, considering (N1)

m̄ txòm, à ò-tā ǁé (My uncle, you are going to die)

the grammatical (future tense) marker ò-tā ǁé (are going to die) is in the process of getting abstracted from a lexical (literal) ú tā ǁé (go and die). Perhaps a little of the literal “go and die” remains in the abstract, future expressing “will die”. It is nevertheless difficult to assess the degree of semantic bleaching. It might be assumed to be proportionate to the coalescence degree, ò-tā being closer to (and thereby potentially more evocative of) the literal form ú tā it derives from than oga is to its putative parent ú kā. (In other words, the latter can be deemed more coalesced than the former, thereby giving less away of its origins.) Yet, isn't present-day English completely insensitive to the literal motion meaning when using be going to as a future marker - while the motion verb (going) is perfectly visible, with no erosion or truncation?

Share Old Norse Linguist

Ultimately, we observe a certain disparity in the morphosyntactic paths followed in the different dialects of !Xun by the development of the grammatical future, and (perhaps) in the degree of its advancement; we can also observe a unanimous convergence of the semantic path: it is systematically a primitive verb of motion, *ú or *gǀè, from which the expression of the future, whatever its form and syntax, derives. Since the youngest (reconstructed) common ancestor has no future marker, this latter similarity cannot be inherited. Heine then formulates the hypothesis of a great tendency, whereby the semantic shift precedes and drives the morphological/syntactic transition -“structural change lags behind semantic change.” (Heine, Kuteva & Narrog, 2017, p.23)

Heine is a specialist in African languages, and the comparative study of their respective dialects is extremely instructive. As it were, dialects synchronize diachrony, bearing witness at the present time to several stages in a similar ongoing process.

Many synchronic differences between dialects can in fact be interpreted as reflecting differing stages or ways of grammaticalization. (Heine & Reh, 1984, p.91)

However, a certain degree of their interdependence cannot be excluded. As well as reciprocal influences through language contact, something - features or tendencies at work - inherent in the inherited linguistic system could explain the similarity of grammaticalization processes originating after divergence from the shared ancestor. (Heine concedes the hypothesis of an inherited shared drift while the process referred to remains nebulous.)

Here, cross-language studies substantiate the demonstration. Staying in Africa, verbs of motion (together with verbs of volition, we remember the English modal auxiliary will) are the most frequent sources of future markers.

• The Nilotic dialects Acholi and Lango resort to a periphrastic construction where bino (to come) is followed by a verbal infinitive form.

Here again, the dialectal comparison freezes unequal degrees of advancement of a (same kind of) grammaticalization. I will eat is an a-bino cammo in Lango and an a-bi-camo in Acholi, the latter having (probably) further coalesced *an a-bino camo into *an a-bino-camo (affixation), later eroded into an a-bi-camo.

• Standard Ewe marks the future by prefixing verbs with á- from the verb vá- (to come).

• Some Kru languages use a derivative of either come or go.

• Duala, a Bantu language, uses derivatives of both to nuance future expression.

Ultimately, the languages just mentioned share with !Xun the same “channels of grammaticalization” - not only the same semantic source, but also the same types of constructions - a complement-based (or particle-based, both differing perhaps only, as we observed, in the degree of coalescence/grammaticalization) and a serializing futures - while a secondary, periphrastic adverbial construction is also observed. (Heine & Reh, 1984, p.131-132) Although contact influences are acknowledged, the respective geographies of the Nilo-Saharan (Nilotic), Niger-Congo (Kru and Bantu) and Kx'a (!Xun) families of languages are mostly disjointed. A genetic explanation for the similitude is also discarded.

But the distance is even greater with Indo-European languages, where the same great tendency can be observed. We remarked earlier on the English be going to. From this entry point, we shall now turn our attention to Germanic cases of grammaticalization of the future tense from verbs of motion.

Subscribe now

African dialects, as we have just noted, offer a snapshot of grammaticalization in progress. As we recall, we can suspect semantic bleaching not to be fully realized in the !Xun (N1)

m̄ txòm, à ò-tā ǁé (my uncle, you are going to die)

where the future tense marker ò-tā (are going to) still bears the form and phonetic of ú tā ǁé (go and die) - and might thereby be evocative of the former.

It is reasonable to assume a similar ambiguity - between motion and future interpretations - for the English be going to + infinitive - an ambiguity that lessens as the phrase enters grammar. (Heine, Kuteva & Narrog, 2017, p.23) Let's observe a passage in 15th-century English (1477, Mubasshir ibn Fatik, Dictes or sayengis of the philosophhres; in Traugott, 2012)

ther passed a theef byfore Alexandre that was goying to be hanged whiche saide …

(a thief who was going to be hanged passed before Alexander and said …)

Here, we visualize clearly the thief literally passing by Alexander on his way to the gallows. But the exact same phrase can also be experienced in a modern way, as an expression of the future stripped of all evocation of motion.

The following passages trace the gradual emergence of the construction in The Penn-Helsinki Parsed Corpus of Early Modern English. Throughout the first half of the 17th century, be going to is mostly followed by a place name, and fewer than ten occurrences by an infinite verb. In the very first we encounter (1599)

Sir, the Germane desires to haue three of your horses : the Duke himselfe will be to morrow at Court, and they are going to meet him.

the preceding clause indicates the place (at the Court) and time (to morrow) of future action (meet him): this action is both “meet” (in the future) and “going (to that place)”. Interestingly, the interpretation can be wholly that of motion without loss of the future tense: to morrow already bears it. In this early instance, be going to + infinitive is a redundant future tense carrier.

Next, a passage by Shakespeare, in the “buck-basket scene” from The Merry Wives of Windsor (1599), catches our attention. (The merry wives trick Falstaff into hiding in the basket, which their servants then carry to dump into the Thames.)

Looke, heere is a basket, if he be of any reasonable stature, he may creepe in heere, and throw fowle linnen vpon him, as if it were going to bucking

A rare occurrence followed by a gerund, which can be interpreted as a place (the place of bucking) or an activity (the process of being bucked, with fowle linnen (foul laundry) as the passive subject).

Just like above, in the following (1647-1648)

It is now about 12 of the clock, Mooneday noone and my Cozin Dalison is going to take water for Gravesend. Shee will bee at Deane Tuesday night.

Mooneday noone already bears the temporal reference, making is going to (take) potentially redundant as a future marker.

In a late 17th century instance (1696)

The greatness of your Necessities, Tam, is the worst Argument in the World for your being patiently heard. I do believe you are going to make me a very good Speech, but

the context involves no indication of place and time, and abstraction to a mere future tense marker has clearly taken off.

What drives language change? What force causes lexical items to shift to grammatical ones? It is usage, it seems, that secretes grammar.

… [interlocutors] constantly propose new discourse options, and some of these new options may be used regularly and give rise to new patterns of grammar.

(Heine, Kuteva & Narrog, 2017, p.23)

Source selection - verbs of motion, for English will, as we have seen, intention - is governed by semantic implication. In to go (to) or to come (to) space and time fuse. Both imply an action that unfolds in the future, namely, until (time) arrival where (place) the verbal complement points to. In short, the scenario might be as follows. A language originally lacks a stable, unambiguous means of expressing some abstract notion - perhaps interlocutors are often forced into periphrases, misunderstood, or asked to repeat. They unconsciously take advantage of a semantic bridge, which usage formalizes. A space-time metaphor.

Once grammaticalization is complete, the frequency of occurrence in the corpus increases. Analysis of discrete samples is no longer adequate to capture patterns. Hilpert (2008) proposes a quantitative study of the Germanic future: he scans corpora for collexemes of future-expressing constructions, the lexical items they statistically attract most. (Co-occurrence frequencies are corrected for total corpus frequencies.)

In the British national Corpus (late 20th century), depend, know, remain and become rank among the top collexemes of will, while be going to strongly attracts do, get, say, put or kill. (Hilpert, 2008, p.40-41) The latter clearly involve an intentional and voluntary agent who has control over an action (agentiveness), and they require or likely come with a direct object (transitivity). In contrast, the former are non-agentive. Arithmetic confirms that the two constructions expressing the future have distinct semantic preferences, which raises questions about their interchangeability. Why does is going to attract agentive, imperfective and transitive verbs? Hilpert's statistics also tell us about the paths of grammaticalization.

Share Old Norse Linguist

Bybee and Pagliuca (1987) confirm the polysemy, in many languages, of future markers. Unsurprisingly, modal nuances abound among the additional meanings. Future tense is a very special tense indeed. Unlike past and present tenses, it concerns what has not yet taken place: realization, at the time of utterance, is no more than a possibility - backed up by a desire, an intention, an obligation, a strong probability. The semantic halo of a future tense marker, the authors recognize, is linked to the meanings of its lexical source. Old English sċeal was primarily used for commands and modern shall still denotes obligation while will is historically desire-based; movement-based futures, for their part, carry less nuances of modality. The authors further suggest that there exist a typical, four-stage semantic progression that holds across languages. Hilpert account for the model as follows:

1. Obligation, desire, ability, come, go

2. Intention, root-possibility, immediate future

3. Future

4. Probability, possibility, imperative, use in complements

It makes two notable hypotheses. First, both modality-based and movement-based futures develop necessarily through an early stage where intention is expressed. (This may not be clear from the outline above, where the second stage seems to present several semantic options including intention, but Hilpert makes otherwise explicit that this has been posited by Bybee and Pagliuca.) Second, future marking and expression are not the term of semantic development. The grammaticalized future marker is the source from which other meanings develop. (Hilpert, 2008, 24-27)

Hilpert puts this model to the test by systematically studying the collexemes of three movement-based future tense marker, Dutch gaan, English be going to and Swedish komma att. (Hilpert, 2008, p.113-122; p.126-131)

While modern be going to, as we have mentioned, is found to attract “verbs that are transitive, punctual, and highly agentive”, verbal complements to modern gaan are most often “intransitive, temporally extended, and non-agentive”. Diachronic corpus studies “suggest that these differences are not due to recent semantic changes”: despite similarities in the broad outlines, the details of their semantic paths differ.

Dutch, like English, readily expresses the future with the present tense, but also features two specific periphrastic constructions: resorting to the auxiliaries zullen and gaan. zullen is a temporal auxiliary: of the future (will, is going to) when conjugated in the present tense, of the conditional (would) when conjugated in the past tense. But it is also a modal auxiliary, with nuances of possibility, probability, moral obligation (probably then closest rendered by the German sollen). The different shades are difficult to disentangle in

Nou, verbeter je maar in het vervolg. Morgen zullen we je wel helpen.

(Well, improve yourself in the future. Tomorrow we will (be there to / be able to / be glad to) help you.)

(Louis Couperus, Noodlot, 1918)

Here, zullen clearly works as a future auxiliary, underpinned by the time adjunct tomorrow, but the context (the addressee is invited to improve, he can still be helped tomorrow, but this may be the last time, and the adverb wel bears a potential hue of likelihood) reawaken the modal flavors.

A modern grammar will endorse gaan (to go) as a future auxiliary

1. when the agent literally is moving to a place to perform an action,

2. more generally, to indicate any kind of intention or plan to perform the action, or

3. to indicate the start of an action in the future. Het gaat zo hard regenen. (It is going to rain / start raining. [It doesn’t rain yet])

and deplore that, in practice, the use of gaan is stepping on zullen's toes.

In the 16th and 17th centuries, verbs of movement and posture are the ones most frequently attracted by gaan. (Hilpert, 2008, p.116)

Nu willic gaan lopen al in mijn huus. (Now I want to go home.)

Daer gaet hij strijcken! (There he’s running off"!)

Besides lopen (walk), strijken (run off), reizen (travel), liggen (lie), treden (step), verhuizen (move), we also find some activity verbs such as preken (preach) or stellen (put). We can note the preeminence of agentiveness, intention, and atelicity (lack of a specific end point). (To preach (preken) is intentional, accomplished by an agent, and continues as long as the action occurs. In contrast, to pronounce is telic - it requires something to be pronounced - a speech -, and concludes by the end of it.)

Subscribe now

Second-period verbs, in the 18th and 19th century extend their span to telic verbs such as opzoeken (find), onderzoeken (analyze), vertellen (tell), geven (give) or nemen (take); intention and agentiveness fade, as for instance gaan sterven (die) joins the top collexemes.

Uilenspiegel zeide tot Nele: -Liefste nu gaan we sterven. (Uilenspiegel said to Nele: -Darling, now we’re going to die.)

The trend extends into modern times, with the notable attraction of cognition and emotion verbs: beminnen (love), voelen (feel), twijfelen (doubt). The latter are low in agentiveness and intention, as the following example clearly shows.

en daardoor gaat ge twijfelen aan het echt-yijn dier (and that makes you doubt the reality of that benevolence)

Whereas gaan has a clear preference for verbs of movement and posture, be going to particularly attracts speech act verbs. Such preferences, Hilpert postulate, drive semantic development. “Verbs such as answer or begin are perfective, and thus prefigure the modern tendency of be going to to express punctual, telic events.” Like gaan, be going to begins (1710-1780) by attracting verbs high in intention and agentiveness: say, fight, give, make. Unintentional, non-agentive future events make their appearance in a second period (1780-1850).

In the true sleepy tone of a Scottish matron when ten o’clock is going to strike.

So far, in spite of the different paths taken, Bybee and Pagliuca's hypothesis holds: intention dominates the early stages. Not so for the Swedish future marker komma att, diachronic statistics reveal: its beginning (16th-18th century) most frequently describes “involuntary reactions and non-agentive human activities”, e.g. förakta (despise), sova (sleep), rodna (blush) or höra (hear). (Hilpert, 2008, p.129-131)

Men man kommer att höra talas om honom! (But we are going to hear people talk about him!)

The next period (19th century) continues to feature involuntary emotional responses with gråta (cry), glädja (delight) or skratta (laugh); yet, intention is on the rise, or at least polysemy opens the door for intentional interpretation. For instance, intention, futurity and modality conflate in

Snart kommer nog andra att säga dig det, mumlade den gråskäggige. (Soon others will tell you so, mumbled the man with the grey beard.)

Only recently can komma att be “felicitously combined” with intentional activities: klara (manage, succeed) or skicka (send).

Ultimately, Hilpert's quantitative analysis succeeds in particular because it combines large number crunching with the pleasing and epistemologically sound examination of discrete corpus occurrences. In Swedish, abstraction from the notion of movement to that of future operates without the intermediary of intention.

There are, however, a few objections to the study of collexemes outlined above. Since large numbers are required, history is necessarily cut into thick slices. The only way to feel the nuances of evolution is still to consult discrete samples that line it up. Hilpert seems interested primarily in the characteristics, in particular the intention and agentivity content, of verbs frequently attracted by our auxiliaries. He compiles frequent collexemes, assesses their degree of intention against the dictionary - weak for sterven (die) or gråta (cry) or skratta (laugh), strong for nemen (take) -, and draws conclusions. Yet the degree of intention is always contextual: the actor intends to laugh and cry. Hilpert himself concedes this incidentally: he justifies an incongruous occurrence of a verb of intention, hämnas (to take revenge) in the first period of komma att by its denaturing by the context into denoting a “reaction to an emotional state”, in reference to

han är missnöjd och han kommer nog att hämnas!

(He’s angry, and he’s probably going to seek revenge!)

While requiring intention, the verb hämnas thus bears some resemblance to the other distinctive collexemes förakta (despise) and rodna (blush) which denote involuntary psychophysical responses. (Hilpert, 2008, p.129)

Finally, both Hilpert and Bybee and Pagliuca may be objected to for the sophistication of their argumentation; to examine whether “future meaning” historically derives from “meaning of intention” in movement-based futures, it is ultimately not necessary to recount the entire history of occurrences, and measure their degree of intention, either by calculation or patient exegesis. An look at the auxiliaries and common sense are enough: to go (gaan, be going to) or to come (komma att) require an agent, who intends to do so. There is always intention in verbs of movement. In their defense, these studies, notwithstanding the theoretical quibbles, are an encouragement to go through the corpora, feel the semantic nuances, and understand the usages, particularly in those of the languages evoked that are foreign to us.

Share Old Norse Linguist

Across languages, future tense develops its markers quite unanimously from verbs of movement or modality. In contrast, the semantic, morphological and syntactic pathways are very diverse. In particular, it is difficult, in view of the arguments put forward earlier, to conclude that there are universal stages of semantic development. Their authors are the scientific grammarians dreamed of by the late 19th century:

Als Aufgabe der wissenschaftlichen Grammatik können wir kurz die Erklärung der Formen einer Sprache bezeichnen und zwar in der Weise, wie sie entstanden sind. (Müller, 1976; in Heine, 1984, p.266)

“We can briefly describe the task of scientific grammar as the elucidation of the forms of a language, namely in the way they have developed.”

The future is a very special tense. Unlike the past and the present, it doesn't really exist: it doesn't exist yet. It is no surprise that it often carries a bit of that modal tone that connotes the real and the possible. Despite the variations, the idea of intention is never far away. It accompanies the early metamorphosis of Dutch gaan and English be going to from the literal to the figurative. It is obvious in desire-based futures, as English will. In all instances, the metaphor is very intuitive: intention wishes present thoughts to become future reality; spatial movement implies and evokes the time needed for the journey.

Language, we say, is in motion, and grammar, usage that has become form. This is rather paradoxical. If grammar is to accurately describe the rule system of a language, those rules and that language must be fixed. (It is impossible to take a snapshot of a body in motion.) If grammar is to prescribe the correct use of a language, its mandate permits no innovation. (Innovation requires - at least infinitesimal - dissidence from existing rules.) We can thereby claim that the grammar of a language in motion doesn't exist. It might be the invention humans infatuated in typology and completeness. The advent of intelligent machines is perhaps proof that, equipped with certain neural networks, one can finely process and generate language without awareness or abstraction of the presumed rules that regulate its usage. Naturally, if the existence of grammar is doubtful, so to is that of grammaticalization: language and its uses just continue their great intuitive movement, where mores, society, technology and communication intermingle.

Subscribe now

References

Heine, B., Kuteva, T., & Narrog, H. (2017). Back again to the future: How to account for directionality in grammatical change. In W. Bisang & A. Malchukov (eds.), Unity and diversity in grammaticalization scenarios (pp. 1-29). Berlin: Language Science Press.

Heine B, Reh M. (1984). Grammaticalization and Reanalysis in African Languages. Hamburg: Buske

Hilpert, M. (2008). Germanic future constructions: A usage-based approach to language change. John Benjamins Publishing Company.

Traugott, Elizabeth C. (2012). Grammatical constructionalization: Rethinking grammaticalization in the light of constructionalization. Paper presented at the University of Santiago de Compostela, 17 October 2012.

Anthony Kroch, Beatrice Santorini, and Lauren Delfs. 2004. The Penn-Helsinki Parsed Corpus of Early Modern English (PPCEME). Department of Linguistics, University of Pennsylvania. CD-ROM, first edition, (http://www.ling.upenn.edu/hist-corpora/).

Bybee, Joan L., & Pagliuca, William. (1987). The evolution of future meaning. In A.G. Ramat, O. Carruba, & G. Bernini (eds.), Papers from the 7th International Conference on Historical Linguistics [Current Issues in Linguistic Theory 48] (pp. 109-122). John Benjamins

Not to use the tools born of the great odyssey of science and technology is, at minimum, to display affectation - the pretense and artificial display of maintaining outdated methods. A ridiculous conservatism, some would argue. With no regard for the cumulative millennia of expended nerves, blood and brain spent developing progressive machines, finally computational ones, whose effects have freed each new generation from the burden of another drudgery. A kind of refractory autism, that of a post-Gutenberg revolution copyist, of a teenager, secret Friday-night video-game player, who in class makes it a point of honor to do all his mental arithmetic in his head.

Subscribe now

Yet, the printing press did not penetrate every cottage, to replace there something. Quite the opposite, it catalyzed the massive training of minds that led to the advent of the computer scientist, the processor-designer, and silicon-based computing centers. It has opened up an ascending movement, where minds and machines, necessarily, progress together, reciprocally perfecting each other.

Generative artificial intelligence ushers in an absolutely new era, which clearly threatens human mental forces with dereliction and desuetude. But how, both intelligences complement each other!, already protest the avant-gardes of the future and the chip peddlers. By using an Archimedean lever, your physical forces are magnified many times over. The result, the total work expended, is incommensurably greater. The strongest of colossi is exceeded by the average stature when equipped with the proper apparatus, thereby depriving the colossus of all comparative advantage, other than decorative, on an outmoded note. There remains, in a special corner of the universe, the Guinness Book of Records, and Olympic contests, a niche market for the antiquarians of physical strength. In short, the lever does indeed complement human force, but also dispenses with any lustre in the matter, except for the display. The utterly unsentimental demands of the social machine and the productive machine - in a word, of progress - ultimately care nothing for the author of the act, as long as the force is deployed. The colossus can recycle itself into mannequin work.

We are inclined to think that the analogy holds entirely when considering the neuronal counterpart of the lever. The average translator endowed with an internet connection is more accurate, more productive and more multilingual than the most erudite of scholars polished by the ages, when it comes to the majority of materials - say, with the minute and shrinking exception of the most sophisticated. The same goes for summarizers, exegetes and quill-drivers. Catchphrases and headlines coiners, office strategists, advice-givers or speech-, ghost- and scriptwriters. As with physical strength, the machine couldn't care less about your name or the glow your remnant instinct for survival and distinction would like to associate with it. If you score very high in numerical and verbal processing, you will soon be impressing no one but yourself. Because your results, which you know are sincerely natural and naturally high, are already by far dwarfed by the minimal, routine efforts of a naturally more robust and infinitely scalable infrastructure.

As with physical force, the great tendency at work, the irresistible drive of progress, prompts our species - that of the consumptive majority or mass society, the infinitesimal top few having almost already branched off onto a new branch of the phylogenetic tree - to accommodate its averaging out, or so to speak, if we subtract from the term its connotation because the point is to argue, rather than polemicize, its mediocritization in intelligence. Yes, human and machine intelligences complement each other, and their conjugated output could very well anonymously surpass any cognitive abilities hitherto conceivable. But this is at the expense of the human share in these capabilities: if you don't use it, you lose it.

Repeat several times if you have to, and remember: if you don't use it, you lose it.

This is, organically, biology. The principles are many millennia-old and now well known to competent neuroscientific services. At the cellular level, a cortex of specialized units is activated by the rhythm of their usage. Discharges are electrical. Electricity runs along the axons, those long wired extensions of nerve cells, whose myelin sheaths make them look just like high-voltage cables. The arithmetical sum of activations on axon endings afferent to a subsequent cell determines signal transmission. At the axon-neuron junction, the electrical signal more precisely converts into a chemical signal, and we reach the molecular level. A set of messenger molecules, the neurotransmitters, trigger the opening of ion channels across the membrane of the recipient cell, and the difference in charge between the latter’s intercellular and intracellular milieu imparts the electrical discharge to it. The recipient cell is activated. Electricity, chemical bridges, electricity, etc. Billions and billions of neurons, their aggregates in lobes and crowns.

Share

We said neuroscience was highly evolved. Despite obvious progress, many mysteries persist: consciousness, emotions, abilities, memory, will - and, of course, language learning. One thing is however certain and known to the lowliest biologist. The more active a brain function, the more efficient it becomes. Conversely, an abandoned circuit is lost to excellence. It falls out of order. The bed of the electric stream narrows into a creek and eventually dries up. (Deshpande & Wang, 2022) The phenomenon or biological capacity underlying this reinforcement or deprecation of circuits is cerebral plasticity. Axons have the power to reorganize their terminations, so as to connect with one recipient neuron or another. This plasticity enables learning, that is, the improvement of a capability, by reinforcing a specific neural circuit. The brain learns Old Norse or, say, arithmetic. But conversely, the exercise of learning improves plasticity: the more the brain learns - of something - the more plastic it becomes, and quicker to learn the same kind of thing, or whatsoever.

In short, practice enhances plasticity, plasticity enhances practice, and so on - we are indeed quite a machine! Understanding a little bit of the mechanics involved already indicates what needs to be done.

But we are wandering. We are giving a long answer to the question you did not and will not ask. Not because it does not cross your mind, but because it doesn't make sense, and will less and less, to ask it. You might still, for a moment, wonder in your inner self who composed, Galois or the machine, but seeking confirmation from the author, the machine, or both (since you would be postulating the possibility that they might be one and the same) would entail being better able to trace the origin of the reply provided, than that of the productions to which the question pertains. In short, a little like emotionally asking a probabilistic generator: are you truly sincere? And the foreseeable progress of machines will further blur the dividing line, your reliability - and that of the machine - to detect it will disappear almost perfectly, thereby finishing off the deprecation of the question - and that of authorship.

Subscribe now

Please note that Galois will not ask their reader if he is human either, and as you understand, they could not care less, since they are in a position to demonstrate to no one the virtuosity, depth, erudition or pertinence by which they might have wished to illustrate themselves. And if they lay down their arms in defense of talents that you might have attributed to them if you were in a position to reliably attribute “their” productions to them, they don't even ask for the benefit of the doubt: they would risk receiving from you, as a substitute, the kind of fair words and patronizing fawning a child receives from slightly embarrassed adults, when he has made a nice drawing, it is clear that he cannot be expected to become a new Edward Munch, but no one wants to be less than encouraging. In truth, anyone who doubts that your works are your own, but hypothesizes, however faintly, that they are produced by calculation, will soon be paying you the quintessential compliment.

Here, a farandole of other objectors, at the opposite end of the spectrum, enters the scene. Word-processing machines will never rival humans! They lack Q, where Q denotes a quality that, in the best-intentioned debates, is still widely regarded as strictly human, such as emotion, creativity, esprit-de-finesse, the je-ne-sais-quoi or its close kin, the almost-nothing. In general, those concerned are told, in retort, that they are proving, at the very least, blindness to facts, and certainly, a poor capacity for anticipation, typically ascribed to a deficit in scientific background. Attempts are made to enlighten them on the workings, prowess and promises of artificial intelligence, but most often to no avail. There always remains a little something spiritual to be defended for these idealists of human preeminence preservation. They worship, deplore our very 21st-century scientists, a brain in a jar of formaldehyde in a decaying natural history museum amidst a smart city rocketing with flying vessels.

In the absence of weights and measures, these conviction-based disputes can go on forever. Yet, we can try to slip in a small but essential accounting element. As we have said earlier, the human being, in the mundane, majority use of their natural intelligence, has already been replaced - or is imminently replaceable subject to a few machine-world interface tweaks. Moreover, innovation is in full swing. The treatment of its disconcerting exponential acceleration is itself already archived among the platitudes. And the confluence of financial flows towards yield-enhancing automation technologies forebodes the perpetuation and reinforcement of the brain drain fueling the trend. A shrinking caste of precision engineers is perfecting at high output, in the machine room, its own robotization.

Thereby, the share of mundane intelligence soars. The movement is twofold. The machine conquers, and thus trivializes, the complex. Humans standardize, average out, in a word, simplify. The machine's democratic accessibility decisively explains the latter. Artificial intelligence, for its part, does penetrate the remotest cottage. The machine humbles us, it is close to being better than us, in most cases. We have it at our fingertips, we overuse it, and our own capabilities recede accordingly. Here again, financial incentives come into play, and they don't encourage exercise, which calls for voluntary impetus from a certain discipline.

This still leaves the ever more marginal upper end of the Gaussian distribution: the virtuoso exercising his predilection might embody that very human something longed for by the aforementioned cantors of some kind of confident humanism. Unlike the mundane performances of intelligence, his spectacle is not subjected to the binary or further discretionary assessments ruling the productive sphere: it would be injurious to pass it by the sieve of Quality of Service. He represents, versus the mass-produced, the limited edition: even if, somewhere down the line, the self-assembly table is to some extent produced by humans, they only intervene insofar as they fulfill their function. They are invariably content with a badge, where the virtuoso had a name, the only thing about him that the merchant sphere - publisher, producer, media or record company - exploited, for want of being able to disentangle and securitize its prowess. In direct line with a very exclusive world of ideas, he used to descend back amongst mortals with an illuminated forehead. Chosen, he was gifted - gifted with a style: he was inimitable. (Here, the anecdotal forgers of history were also virtuosos in their own way.)

Clearly, this hypothesis of unequaled human virtuosity - it suffices that it be expressed by a few, Bach, Cicero, or Hieronymus Bosch - coincides exactly with the disputes over mind-body dualism. If we believe that perfect reductionism, of soul to brain, can ever be mapped, there is nothing but computation up there. Otherwise, we have a soul supplement. The soul is the organ of feeling, inspiration and meaning. It has a great deal of intuition, but is obviously unable to explain positively how it understands what it is saying. Yet it does, while the statistical machine spews out.

The total replacement of human intelligence by the machine therefore has the same debate-producing potential as God, eternal life and the gender of angels.

Subscribe now

Not using intelligent machines in every instance requires, thus, discipline. But discipline, like beauty or professionalism, admires itself. After a while, one finds pleasure in it. Is Galois, then, a narcissus of self-perfection? Halfway between an ascetic and one who delights in well-turned phrases? Perhaps a little, like all of their era.

You seem to persist in asking yourself the question you won't ask them, that of why Galois bothers to compose and study with their five senses - Galois still regularly writes with a fountain pen - well, at least, sight, hearing and touch. As we said, their motivation here is by no means the pursuit of results. For their slow, too-human, too-fallible processes would be perfectly counterproductive to that end. Reciprocally, the result-driven motivation, the one that rushes to the automatic translator or answering machine, takes on a productivist perspective: maximizing speed, crossing the quality threshold, fitting into the standard gauge.

Even if it is impossible to know, you can always bet on having a virtuoso within yourself, a human part. In any case, you can indulge in honoring yourself as an end, rather than diminishing into a means. Once extricated from the realm of accountancy, the motives and criteria of your achievements have your own worldview as their yardstick. Galois is their own standard.

Is Galois a Luddite, a helmsman of progress, an enemy of the Enlightenment? If they managed to arouse such suspicions, they stand corrected. Galois marvels at the technologies their time is perfecting. Of scientific pedigree, they are a fervent contemplator of algorithmic elegance.

What they maintain, however, is that they resist, in all situations requiring computational outlay, using the machine in the first place. They first try to find the solution by themselves, feeling their way through the darkness for a while if necessary. The machine can, of course, solve the same problem in an instant. But they would be depriving themselves of the exercise. They remember the well-stocked graduate mathematics textbooks. You don’t learn here by reading the solutions, you learn by searching. By searching for a long time. This is where the brain’s microcircuits are refined: you compare and contrast, you develop patterns, you progressively master the algebraic material. And they have learned their lesson.

When they want to write in a foreign language, Galois begins by flexing their strength, giving their best, most genuine effort over an extended period of time, until they achieve a certain satisfaction. Then, they might turn to the machine and converse. They convert back and forth between languages, polish at length and make the final judgment. Yes, in a way, Galois derives time and energy from the great general machinery. They divert it to drive their own little inner theater.

But you learn Old Norse, you are a little like them.

Share Old Norse Linguist

References

Deshpande, Gopikrishna & Wang, Yun. (2022). Noninvasive Characterization of Functional Pathways in Layer-Specific Microcircuits of the Human Brain Using 7T fMRI. Brain Sciences. 12. 1361. 10.3390/brainsci12101361.

Grammaticalization, in the broadest sense and abstracting for now from terminology debates, can be defined as the process which, in the history of a language, progressively renders the expression of a notion grammatical.

How is such a notion originally rendered, if not grammatically? Looking at modern English by way of illustration, it seems that the notion of definiteness is grammaticalized, whereas the notion of size is not. The distinction between indefinite (I saw a dog) and definite (the dog seems happy) is carried by grammatical items, namely indefinite articles (a/an in the singular, absence of an article in the plural) or definite articles (the). In contrast, we have to resort to the lexicon to express something pertaining to the size of the dog. Here, the choice is abundant and unregulated, both register and form. It extends to the spectrum from slang to preciosity, archaisms to neologisms or loanwords. It is not limited to any one grammatical category or construct - adjectives, adverbs, complex periphrases and whole clauses have full license.

A colossal dog stands in my way.

The dog dwarfs the cat.

Comparison, metaphor and other stylistic devices further extend the range of possibilities, when what is spoken of gives way to an image bound only by the limits of the imagination. In Lewis Carroll, it is the rat-hole that defines the dimensions of the small passage, after Alice's fall through the rabbit-hole.

Alice opened the door and found that it led into a small passage, not much larger than a rat-hole.

Subscribe now

Here is how Alice pictures the notion of shrinking in size, in order to access the small passage.

"Oh, how I wish I could shut up like a telescope! I think I could, if I only knew how to begin."

A wish that later comes true.

“What a curious feeling! […] I must be shutting up like a telescope!”

In the latter sentence, must be is the grammatical vehicle for the expression, thereby (at least partly) grammaticalized, of belief or knowledge, and more broadly, of epistemic modality; modality for the expression of the possibility or necessity for the reality to be a certain way, of epistemic flavor because the possibility is assessed based on some body of knowledge, here empirical (“What a curious feeling!”). Might, may, must belong here. Modality has deontic flavor when is expressed the necessity or degree of freedom for reality to be otherwise: reality has to or is allowed to be transformed to accord with certain rules, norms or desires. The injunctive must belongs here. The flavor of modality is dynamic when the possibility is contingent to some qualities, capabilities or circumstances. For instance, the dynamically modal could - a matter of aptitude - in

“Oh, how I wish I could shut up like a telescope!”

Share Old Norse Linguist

The modal verbs we have just reviewed are at home in English grammar. It is legitimate to wonder why - why these are more grammatical than other words, maybe small or telescope - in other words, precisely what makes them grammatical items.

English modal verbs are of a very special kind. They are auxiliary verbs, verbs that, in accordance with etymology, help others to express a notion, more than they themselves carry meaning. They qualify for the category under the sieve of certain criteria, of which we detail briefly a few. First, auxiliaries can be negated by not, while lexical verbs can’t. They further have negative inflected forms, as in He can’t be here before late. Second, they take the first position (inversion with the subject) in an interrogative clause. Must she wear a coat? Third, they can encode a complete verb seen previously. The second could in Alice’s

"Oh, how I wish I could shut up like a telescope! I think I could, […]”

picks up the first complete I could shut up (like a telescope).

Furthermore, they qualify as modal auxiliary verbs, a subset of the former group. Further criteria apply. Contrary to non-modal auxiliary, they show no subject verb agreement. This might happen and not *might-s. They lack untensed forms, namely the unmarked verb, such as be in I’ll be here, and the gerund-participle, such as being in He dislikes being bullied. *He will must make do or *He was canning speak Old English are clearly grammatically wrong. They can take a bare infinitival clause (a clause with an unmarked verb) as their complement. We have

I could shut up like a telescope!

where the bare infinitival clause is highlighted, while the following sentences are ungrammatical:

*I could to shut up like a telescope
*I could shutting up like a telescope

In short, modal verbs constitute a restricted set of words obeying their own rules, and far less bearers of meaning than means to a function.

Subscribe now

So far, we have merely clarified the demarcation line between grammatical elements, function words, and lexical elements, content words. Some words belong in the grammar book, others are confined to the dictionary. Yet, we have said nothing about entering grammar: grammaticalization, we claimed, is a process. Let's look at some modal verbs in Shakespeare. A Midsummer Night's Dream was premiered in 1595 or 1596.

Obe. That very time I saw, but thou couldst not,
Flying between the cold moon and the earth,
Cupid all arm'd

The Tempest in 1610 or 1611

Ste. How now shall this be compassed? Canst
thou bring me to the party?

and

Cal. Within this half hour will he be asleep;
Wilt thou destroy him then?

Early Modern English distinguished between a second-person singular (thou) and a second-person plural (you or ye), with verbs agreeing with a distinctly singular second-person subject (distinctly insofar as a second-person plural form of the same verb might exist). All verbs are concerned, including, as we have just noticed, modal auxiliaries. Again in The Tempest

Ste. Give me thy hand: I am sorry I beat
thee; but, while thou livest, keep a good tongue
in thy head.

In Shakespeare's time, modal verbs were more akin to lexical verbs: they followed the then-current pattern of second-person singular agreement. Lewis Carroll's modal auxiliary morphology, for its part, has departed from the standard conjugation table: it doesn't comply with the modern agreement of lexical verbs, which marks (with an -s suffix) the third person singular.

Such morphological reduction - the alienation of auxiliaries from their original category, that of verbs in general, through the loss of morphological features specific to this category - constitutes, according to linguists, one of the frequent mechanisms along grammaticalization pathways. At this stage, our opening definition of grammaticalization requires revision. We have said, in the broad sense, the linguistic process whereby the expression of a notion is rendered grammatical. In a narrower sense, this evolution is more precisely carried by that of a specific linguistic item, from lexical to grammatical. School debates notwithstanding, this is the idea towards which theorists seem to be converging. Notably, Jerzy Kurylowicz: (Kuryłowicz, 1965)

Grammaticalization consists in the increase of the range of a morpheme advancing from a lexical to a grammatical or from a less grammatical to a more grammatical status, e.g. from a derivative formant to an inflectional one.

An elegant definition is that of James A. Matisoff. Grammaticalization is “a subtype of metaphor”, “a metaphorical shift toward the abstract”. Metaphor, he recalls, means etymologically “carrying beyond”, and can be defined as “an originally conscious or voluntary shift in a word's meaning because of some perceived similarity”. (Matisoff, 1991)

Subscribe now

Let's retrace the metaphor for some of our English modal verbs, and get closer to the notion that ultimately interests us: the expression of the future. Here is the drift from Proto-Germanic into Old English and then Middle English for the ancestor of Present Day English to will.

PG *wiljaną (to want) > OE willan (to want, to intend, to desire, to be willing) > ME willen > PDE will

willen still carries the nuances of its parent willan, but its function as an auxiliary has emerged. The verb will, in modern dictionaries such as Merriam-Webster.com, is first defined as the auxiliary of futurity. The Wiktionary, for its part, says “Used to express the future tense, sometimes with an implication of volition or determination when used in the first person”. Intention or willingness contain the abstract notion of future. To intend is to decide, anticipate and prepare for the action to come; willingness desires or envisions it. The lexical meanings of willen fade away, the underlying notion come to prevail: we clearly observe Matisoff’s “metaphorical shift toward the abstract”.

Share Old Norse Linguist

A perspicacious mind will nevertheless have noticed a flaw in the flow of our remarks. We declared our intention to study the grammaticalization of modal verbs, and find ourselves talking, not about modality, but futurity. In fact, languages often conflate the notions of tense (point in time), aspect (extension in time), and mood (modality), making them difficult to disentangle. The modern auxiliary will, to keep with our example, remains polysemic: the context drives the necessary interpretation. Less frequent uses, Merriam-Webster.com tells us, express probability or capability/sufficiency. In Alice's Adventures in Wonderland, “The Queen’s Croquet Ground”, we read

“No, I didn't,” said Alice: “I don't think it's at all a pity. I said 'What for?'“

“She boxed the Queen's ears— “the Rabbit began. Alice gave a little scream of laughter. “Oh, hush!” the Rabbit whispered in a frightened tone. “The Queen will hear you! You see she came rather late, and the Queen said—[…]”

Here, the Queen “will hear” contingently on a future action, namely that Alice does not remain silent. Therefore, will is a clear operator of the expression of the future. Yet it is only probable that “The Queen will hear”. This depends in particular on Alice's behavior and the Queen's hearing acuity. The nuances of tense and mood intermingle to color an expression that only personal interpretation can resolve.

Share Old Norse Linguist

Theorists have examined at length grand tendencies in grammaticalization: firstly, they record a certain universality of its generic morphological, semantic, phonological or syntactic mechanisms; the unidirectionality of their deployment, from the lexical to the grammatical and its strengthening, is a debated hypothesis we shall evoke shortly. Secondly, a certain universality, across languages, of the notions that become grammaticalized (e.g., modality or future tense) and their specific semantic (some say cognitive) pathways: frequent semantic source-target pairs, from a lexical meaning to its metaphorical abstraction, emerge (e.g., from intention or willingness to futurity).

Subscribe now

Grammaticalization is a theory, an idea substantiated by empirical evidence. It has its adherents, and more cautious contributors. It may be judicious here to remark on its premises. Grammar, according to grammaticalization, is a “product of change”: it can “be understood as a historical process embedded in use rather than as a purely abstract synchronic state.” (Hopper, 1996) The phrase “embedded in use” is worth noting here. In the course of its history and use, language is, so to speak, in gestation of its systematisms. Universal mechanisms shape its progression along known semantic pathways. The language gradually emerges from the primitiveness of its fluctuation, from a certain looseness, towards a kind of grammaticalized, ideally ordered state. Based on these premises, the discipline seems to be longing for a large, dual repertoire: that of global mechanisms and that of source-target semantic pairs. Large-scale empirical surveys, such as those carried out by the Africanist Heine (Heine & Reh, 1984), are gathering evidence. Some studies seek to circumscribe and relativize its relevance, without denying it. Ultimately, we can expect to refine our understanding of a language by examining it through the prism of grammaticalization and the encyclopedia of its great tendencies; and we can hope to extract from the universal linguistic pathways recorded truths about human cognition. (Hopper, 1996, p.224-225) In our view, the approach has an additional virtue. By examining the history of forms and meanings, through a kind of etymology of notions, it allows us to appreciate more intensely the polysemy of language, for literary or generally creative purposes.

When meditating on English modal verbs, we have already come across one of the frequent mechanisms on the path to grammaticalization, regardless of the notion at stake - namely, morphological reduction. In Present-day English, the agreement of the modal auxiliary has lost its similarity to that of ordinary verbs. In Shakespeare, we still read

Obe. That very time I saw, but thou couldst not,
Flying between the cold moon and the earth,
Cupid all arm'd

where the modal auxiliary shares the second-person singular (thou) marker (-st) then carried by all verbs.

Share

In his systematic compilation, Hopper (1996, p.223-225) summarizes seven further mechanisms, as highlighted by Heine & Reh (1984):

1. Loss in “in semantic complexity, functional significance, and/or expressive value”. Some say semantic bleaching, and we have taken a long look at how the meanings of to will fade.

2. Loss in “pragmatic and [gain] in syntactic significance”. For illustration, the phrase the fact that traditionally requires the subsequent clause to be true (pragmatically, it is a fact). Recently, this condition seems to be weakening, for example in My opponent has charged me with the fact that I used illegal campaign funds.

3. Reduction in the “number of members belonging to the same morphosyntactic paradigm”. French negation could resort to numerous reinforcers, of which only pas, and more archaically, point, remain.

4. Decrease in “syntactic variability”, that is, “position in the clause becomes fixed”. For example, full flexibility in word order for a Latin sentence, less so for its descendant in Old and Middle French, and none at all for the French equivalent.

5. Increase in “obligatory [character] in some contexts and ungrammatical [character] in others”. Tense, aspect and modality are difficult to express without auxiliaries and conjugation tables. The periphrastic bypass, for example be able to for can, supplies the modal verb's defective forms (namely untensed forms: plain verb form, *I will can come, to-infinitive, *to can, and gerund-participle, *canning) and can therefore be seen as quasi-modal.

6. Semantic, morphosyntactic, and phonetic coalescence with other units. In English, I am going to take the morning flight to Uppsala becomes in some registers I’m gonna take… or even I ma take…

7. Loss in “phonetic substance”. (Phonetic erosion applies to the linguistic item itself, not to its coalescence (6) with other units.) For instance, the Old English negator ná wiht (to be compared with the modern no way) follows the path ná wiht > nówiht or nówuht > noght > n’t.

Share Old Norse Linguist

We have become acquainted with grammaticalization as a linguistic process that gradually metamorphoses lexical items into grammatical ones, following established pathways of language evolution. We have, by way of illustration, visited modal verbs from Old to Modern English, and evoked the frequent intermingling of the expression of tense and modality. The future tense auxiliary will, in particular, bears also modal nuances. It has been grammaticalized as a future tense marker by analogy with its predecessor's lexical meaning - intention and willingness.

Another typical pathway in the grammaticalization of the future tense originates from verbs of movement. This is what we shall soon be looking at more closely, through a journey across several languages and continents - we will evoke the Germanic spectrum, and this will lead us in part three to a study of the future tense in Old Norse.

References

Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J. (1985). A Comprehensive Grammar of the English Language. London: Longman.

Views on grammaticalization. In Grammaticalization. Wikipedia. Retrieved January 18, 2025, from https://en.wikipedia.org/wiki/Grammaticalization#Views_on_grammaticalization

Kuryłowicz, J. (1965). The Evolution of Grammatical Categories. Diogenes, 13(51), 55-71. https://doi.org/10.1177/039219216501305105

Matisoff, J. A. (1991). Areal and universal dimensions of grammaticalization in Lahu. In E. C. Traugott & B. Heine (Eds.), Approaches to Grammaticalization.

Merriam-Webster.com dictionary. Retrieved January 18, 2025, from https://www.merriam-webster.com

Heine B, Reh M. (1984). Grammaticalization and Reanalysis in African Languages. Hamburg: Buske

The present essay aims once again at extrapolating some grammatical truths from the intensive reading of a short text. We shall consider the following sentence.

En þá varð fǫr manna mikil mjǫk út hingat ór Norvegi, til þess unz konungrinn Haraldr bannaði, af því at honum þótti landauðn nema.

Norwegian royalty fears demographic hemorrhage.

“Then there was a very great migration of people out here from Norway, until King Harald forbade it, because he thought the land was becoming depopulated.”

The grammatical mysteries to be solved are many, and we shall focus now on the intriguing ending.

Subscribe now

We highlight the last clause

En þá varð fǫr manna mikil mjǫk út hingat ór Norvegi, til þess unz konungrinn Haraldr bannaði, af því at honum þótti landauðn nema.

that we translated earlier as

“…, because he thought the land was becoming depopulated.”

Here, we cannot short-circuit the entire grammatical analysis; it will lead us to the point we want to emphasize. We already met af því at earlier, a compound subordinating conjunction, subordinating because it introduces the highlighted clause, subordinate to the principal. If we unpack it, we find a prepositional phrase, af því, where því the singular neuter dative of the demonstrative pronoun. We are not surprised by the dative, called by the preposition at, which itself means from, out of, because of. We then find at, which can also function independently as a subordinating conjunction, often to be translated by that. Putting it all together, it gives literally “from this, that” or “because of this, that”.

Our demonstrative pronoun því acts as a cataphoric reference here: it points forward to the “gist” or “content” of the causal adverbial clause the compound subordinating conjunction introduces, i.e. “he thought the land was…”. We might just as well paraphrase as “because of the fact, that...” What is the grammatical function of that here? We might be inclined to think of it as a relativizer, a relative pronoun referring to the substantive fact. This cannot be the case, however, since the clause it introduces is no relative clause, insofar as it does not modifies a noun phrase: it does not further characterize the substantive fact, so the latter could exist without the that-clause. Compare with a nominal phrase featuring a relative proposition: “the cat that sleeps on the mat”. Here, “that sleeps on the mat” characterizes “the cat”. The full sentence

I can see the cat that sleeps on the mat.

can perfectly survive without the relative clause:

I can see the cat.

It is of course not the case with fact in our slightly rephrased clause from Íslendingabók

“…, because of the fact that he thought the land was becoming depopulated.”

since

“…, because of the fact.”

alone is meaningless. We lack the content of the fact, what the fact actually is. The clause “that he thought the land was becoming depopulated” is too essential to be a relative. It is a so-called content clause, or dependent clause, i.e., providing the content that is implied or commented upon by the main clause.

Let's illustrate with further examples where that is no relativizer, but a complementizer, a subordinating conjunction introducing a content clause. In the following sentences

He believes that global warming is not happening.

She is delighted that John got fired.

They wished that the forest fire had been stamped out earlier.

the that-clause functions clearly as a direct object to the verbal forms, respectively believes, wished and is delighted. We can rephrase slightly: in these reformulations, the that-clause logically represents what exactly the subject is, what it implies, its content.

What he believes is that global warming is not happening. (What he believes = that global warming is not happening)

What she is delighted of is that John got fired. (What she is delighted of = that John got fired)

What they wished is that the forest fire had been stamped out earlier. (What they wished = that the forest fire had been stamped out earlier)

In the syntax of those reformulations, the that-clause functions as a nominative predicative, or subject complement, in exactly the same way as “a lone wolf” in “John is a lone wolf.”.

In contrast, for our “cat that sleeps on the mat”, “that sleeps on the mat” is not the content of the “cat”. It just adds a bit more information about it. The same cat can lead a peaceful existence on the sofa or in the kitchen.

So far, our content clauses complete verbs of cognition (to believe, to wish). They can also complete verbs of reporting (to tell) or perception (to hear). They can also very well complement noun phrases, as in

the idea that the Earth is flat…

which in fact we can be rephrased awkwardly for the sake of making a verb (of cognition) appear

what one thinks, that the Earth is flat… or what one has an idea of, that the Earth is flat…

Just like earlier, “that the Earth is flat” is exactly (the content of) “what one thinks”. Back to our Íslendingabók, in

“…, because of the fact that he thought the land was becoming depopulated.”

the that-clause now appears clearly as a dependent clause making explicit the content of the fact, of því in af því at. But why bother? Well, in “the cat that sleeps on the mat”, “the cat” is the logical subject to “sleeps on the mat”. In

…, af því at honum þótti landauðn nema.

því is no logical subject to “at honum þótti landauðn nema” - and in fact “at honum þótti landauðn nema” has no logical subject in the Old Norse sentence.

Share Old Norse Linguist

Let's digress briefly to the virtues of grammar, a matter we shall explore at greater length later on. Jargon, in this instance, is less important than the logic at work. The former is contingent on the grammatical theory to which we refer, and as in any academic discipline, debates, opposing models and grand ideas abound. Rather, the aim is to grasp the overall articulation of a language, and the system of rules governing it. In this way, the concepts of relativizer, complementizer or content clause are mainly manipulated by modern approaches to grammar, such as generative grammar which builds on assumptions we shall come back to. You may not remember them from early school days. A more traditional approach would have said relative pronoun or subordinating conjunction. Nevertheless, the relative pronoun is also, in its own way, a subordinating conjunction, so the relativizer-complementizer distinction seems a sensible way of capturing a nuance in usage.

Accurately mapping the grammatical structure of a sentence is the catalyst for accurate translation and comprehension. Ideally, grammatical analysis should become second nature. This is particularly true for a language with such a rich inflectional system as Old Norse. As we discussed much earlier, considerable variation in word order is allowed, since the grammatical function of words is already marked by case endings: rigidly assigning a place in the sentence to each function becomes redundant with form. Anglophone reflexes for parsing sentences, for spotting subject, verb, complement do not hold up.

As far as we are concerned, knowing that our subordinating conjunction does not act as a relativizer, and in particular does not refer to a logical subject to what follows, is important for grasping the first notion of grammar that interests us today.

Subscribe now

Haugen talks about verb frames. (Haugen, 2015, p.132)

Verbrammene fortel kor mange nominale ledd verbet knyter til seg, dvs. subjekt og objekt (som det kan være flere av).

“Verb frames tell you how many nominal clauses the verb attaches to itself, i.e. subject and object (of which there may be several).”

We are interested in þótti in

…, af því at hónum þótti landauðn nema.

It is the third-person singular past indicative of þykkja (to think). By all means regular, a weak verb of the telja-class. As we have discussed at length, and paradoxically, our Old Norse verb thought has no logical subject. Its verb frame is Verb+Dative according to Haugen’s typology. Our dative here is hónum, the dative singular masculine form of the third person personal pronoun hann (he). The dative is typically the case of the indirect object. Such that “hónum þótti” equates closely to the broken English “*it thought to him”, or closer to correctness “it seemed to him”. “The thought came to him, that…”.

Verbalet tek eit objekt i dativ, men ikkje noko subjekt.

“The verb takes an object in the dative, but not a subject.”

A number of verbs fall into this category, Haugen tells us: hitna (to become hot), lika (to like), litask (to look about), svala (to cool), sýnask (to appear), versna (to worsen), and our verb of main interest, þykkja (to think).

In Norwegian Bokmål, the impersonal construct with dative experiencer is replaced is those instances by a personal construct with (nominative) explicit subject. In particular, our impersonal turn “hónum þótti” is well rendered by “han syntes”, that English can translate into “he feels”, “he thinks”, or (making our dative reappear) “it seems to him”. Nevertheless, there is something special about the verb å synes. A deponent verb, it is a passive form with an active meaning, an exemplar of the famous Norwegian -s passive. We therefore still have a memory of the ancient passive.

In fact, å synes derives naturally from Old Norse sýnask, which we have just listed among the subjectless verbs with dative. Why translate þykkja into Norwegian Bokmål by means of a descendant of sýnask? Well, Norwegian Bokmål does not record, it seems, any descendants of þykkja. Instead, its presumed Proto-Germanic ancestor, *þunkijaną, produced a West-Proto-Germanic cousin from whom the Shakespearean methinks descends.

“The lady doth protest too much, methinks.” (Hamlet)

“Another moon; but O! methinks how slow
This old moon wanes” (A Midsummer Night's Dream)

“Methinks I have
a great desire to a bottle of hay: good hay, sweet
hay, hath no fellow.” (A Midsummer Night's Dream)

Today an archaism, methinks interests us as an English vestige of this impersonal form plus dative. Methinks is better rendered by “it seems to me”, “it occurs to me”, than the simple “I think”, a slight nuance of a passive. Me- is the trace of the old dative, while the third person singular form with its -s ending is clearly impersonal. while the third person singular marked by the -s ending in absence of subject denotes the impersonal form. Modern German also has its archaic descendant, dünken, which the Duden dictionary explains as “jemandem so vorkommen, scheinen”, quite literally “to seem/appear to someone”. Vorkommen in particular renders the passive character: thoughts or impressions come (kommen) to someone. Here, the experiencer is in the dative or, more frequently, in the accusative, the case of the direct object. “Mich dünkt”, where the accusative can be rendered by (the incorrect) “*it seems me”, while a dative-like construct is more natural to English, “it seems to me”.

Share Old Norse Linguist

Now let's look at the end of our clause, which falls under what Haugen calls omformingar (reformulation), more precisely in the småsetningar (short sentences) and, further, the nominativ med infinitiv (nominatif with infinitive) categories. (Haugen, 2015, p.153)

…, af því at hónum þótti landauðn nema.

landauðn is our nominative, i.e., subject case. It is a feminine substantive composed of the substantives land (land) and auðn (wilderness, desert). The whole is given by The Cleasby & Vigfusson Dictionary the English equivalent “laying a land (country) waste”, in other words, depopulation. nema is our infinitive, of a strong verb, meaning to take as the Old English niman (still alive in numb or nim) was superseded by the Scandinavian taka. (In modern Scandinavian we have the Norwegian Nynorsk taka, the Swedish taga, or the Danish tage). The meaning of nema is rarely limited to the literal, and to take sets solely the big orientation. To take away, to bereave of something. To take by force, to seize upon. Accompanied by við, to strike against so as to stop, to come no further. To reach, to touch. To stop. With staðar, to take place. Cognitively, to take in a mental sense or by the senses, to perceive, just like the Latin apprehendere, thereby also, to learn. But luckily the same Cleasby & Vigfusson Dictionary, always rich in literary examples taken from the sagas, mentions our passage from Íslendingabók. Entry nema 2.A.6:

6. to amount to, be equivalent to; ef eigi nemr kúgildi, Grág. ii. 233; honum þótti landauðn nema, Íb. 4; það nemr öngu, litlu, miklu, it is of no, of small, of great importance; bríkr þær er greyping hefir numit, Gþl. 345.

Our infinitive group takes on its meaning: to amount to depopulation, to take the level of depopulation. Or a maybe more natural expression, to become depopulated.

Within the short sentences category, the nominativ med infinitiv configuration is far less common than its twin, the akkusativ med infinitiv. In both cases, it is the reformulation of a subordinate proposition (with a finite, i.e. conjugated, verb) into an infinitive proposition when brought up into the main proposition. The grammatical subject of the subordinate clause before reformulation is switched to the accusative in one case, kept in the nominative in the other. The nominative-preserving construction occurs, Haugen tells us, solely with the verbs sýnask and þykkja, which we now know well. It is precisely their natural construction without grammatical subject, and with experiencer in the dative (“hónum þótti”), that governs the nominative (“landauðn nema”) of their subordinate's reformulation. Haugen indeed says:

Ettersom det ikkje er noko subject i nominativ i samband med slike verb, vil eit subject i ei underliggjande setning bli ståande uendra i nominativ dersom den underliggjande setninga blir løfta opp i den overordna setninga.

“Since there is no subject in the nominative case accompanying such verbs [sýnask and þykkja], the subject of a dependent clause will remain unchanged in the nominative case if the latter is lifted up into the main clause.”

The “place of the nominative” is free in the sentence's inflectional organization, so the transformation needs not resort to the accusative.

Let's take a closer look at the transformation in question. The arrival point is the clause

…, af því at hónum þótti landauðn nema.

which very literally (improperly) translates into

*…, because he thought the land to become depopulated.

This is improper in English, since the verb to think, in its first meaning, “to have a particular idea or opinion about something/somebody; to believe something” (Oxford Learner’s dictionary), requires a verbal direct object to be a that-clause with finite verb. (We can have He thinks to change the tire. expressing intention, or the passive construct They are thought to have exceeded the speed limit.)

We can correct our literal translation into

…, because he thought the land was becoming depopulated.

making the that-clause obvious. It is the Old Norse equivalent of this that-clause that is the starting point of the transformation-according-to-Haugen. Nema is conjugated in the third person singular (in agreement with landauðn) past indicative, nam. We have already the nominative landauðn, which the transformation is to leave unchanged.

Here is our transformation at work.

(at) landauðn nam (that the land was becoming depopulated) →landauðn nema

We shall come back more systematically to the two types of reformulation, accusativ med infinitiv and nominativ med infinitiv in further creative pieces. Yet, let us note here that these have numerous equivalents in the broad Germanic spectrum, and elsewhere. The English to think admits no direct infinitive complement, but consider to wish or to expect.

He wished that the land would become depopulated.→He wished the land to become depopulated.

She expected that the bus would depart earlier.→She expected the bus to depart earlier.

Or more Shakespearean.

Ros. She desires to speak with you in her
closet ere you go to bed. (The Tragedy of Hamlet)

German has infinitive sentences (Infinitivsätze), which can take on subject or complement functions, whose infinitive is accompanied by zu and lacks a subject.

Er behauptet, die Behörden auf die Untat aufmerksam gemacht zu haben. (He claims to have alerted the authorities to the outrage.)

The logical subject of the infinitive is the (non repeated) one of the principal, er, i.e. he. The case of “die Behörden” (the authorities) is the accusative, direct object to our past infinitive “gemacht (zu) haben”. It reformulates the following (uncommon) sentence where dass is the subordinating conjunction (that)

Er behauptet, dass er die Behörden auf die Untat aufmerksam gemacht hat. (literally, He claims that he has alerted the authorities to the outrage.)

And the Norwegian structure is not far off at all.

Han hevder å ha gjort myndighetene oppmerksom på overtredelsen. (He claims to have alerted the authorities to the outrage.)

In German or Norwegian, a verb of perception, such as to see or to hear, is followed by a bare infinitive (without å or zu respectively).

Jeg kan høre toget komme i det fjerne. (I can hear the train coming in the distance.)

Ich höre den Zug in der Ferne kommen. (I can hear the train coming in the distance.)

Subscribe now

We spoke a little Nynorsk (literally, New Norwegian) through Haugen's grammar, which was written in that language. We shall soon explore the origins of Norway's two official languages, Norwegian Bokmål and Nynorsk, compare their distinctive characteristics, and examine their place in the North Germanic spectrum.

References

Haugen, O. E. (2015). Norrøn grammatikk i hovuddrag. Novus.

Cleasby, Richard & Gudbrand Vigfusson. "nema." An Icelandic-English Dictionary. Oxford: Clarendon Press, 1874. https://cleasby-vigfusson-dictionary.vercel.app/word/nema

Old Norse largely preserves the realization /ð/, but the sound is, as we have patiently noted on a few examples, amply "contaminated" by the phonological environment. It devoices into a voiceless fricative, /t/, after a voiceless consonant (notably). It hardens into a voiced plosive, /d/, when it follows a nasal sound, i.e. /m/ or /n/ or the sound /l/.

Subscribe now

Many of the phonological processes at work in the progression from Proto-Indo-European to Proto-Germanic, and on to and during the life of Old Norse, are such contaminations. Grimm's law and its successor, Verner's law, ultimately formalize some of them. Contamination - or assimilation, as grammarians such as Haugen and Spurkland would say. We shall take up and synthesize our proud findings on dental suffixes, following this time Spurkland’s lesson.

As a reminder, our phonetic table of Old Norse consonants. We highlight our dentals of interest. (In Old Norse, the sounds /d/, /t/, and /ð/ are bijectively spelled <d>, <t>, and <ð>, so we can interchangeably refer to these three letters/sounds as the sound or the letter.)

The dental suffix, Spurkland tells us, is primarily realized as /ð/ in Old Norse. (He says, apical suffix.) This is particularly the case when the weak verb is formed using a linking vowel <a>, i.e. weak verbs of the first class, also know as kasta-class.

Among the pasts we have collected earlier in The Saga of Sigurd the Crusader, Eystein and Olaf, let's choose þakka (to thank, to show gratitude) for illustration.

… ok þakkaði várum dróttni feginsamliga …

þakka is our infinitive, which coincides also the present third person plural. Let's recall our representation of weak verbs by their essential forms.

infinitive - (present indicative) - preterite indicative - (preterite subjunctive) - supinum

We then have

þakka - (þakkar) - þakkaði - (þakkaði) - þakkat

where we have highlighted the linking <a> and our voiced fricative dental suffix.

Verbs in classes two and three, namely telja-class and dǿma-class, have no linking vowel, and here assimilations come into play. (Spurkland, 1989, p.100)

det som kan skje, er assimilasjoner mellom bøyningsending and rotkonsonant

i.e., “what can occur is assimilations between inflectional ending and root consonant.”

Let's start by looking at one of our collected verbs whose root ends in a consonant, gipta (to marry, transitive or later, reflexive). We have

gipta - (giptar) - gipti - (gipti) - gipt

(Note in passing the supine gipt, very close to the Norwegian gift, married. From gipt to gift, the /p/, a voiced plosive, has turned into a voiced fricative, /f/, which closely evokes Grimm's law. The phonological laws or processes that Ringe and linguistics identify are at work across the entire Germanic spectrum, major trends that persist as the language evolves, here from Old Norse to its most modern descendant.)

Share Old Norse Linguist

The root gipt- ends with a <t>. In the past indicative, we add to the root the dental suffix and the 3rd. person singular past ending, gipt-ð-i which transforms along

gipt-ð-i →gipt-t-i (assimilation)→gipti

In the second stage, our progressive (the preceding sound influences the following sound) assimilation (a kind of contamination) takes place: the unvoiced plosive /t/ of the root transforms the /ð/ into /t/.

In the third stage, some shrinking occurs: the long (geminate) consonant, here <tt>, is shortened when preceded by a consonant. This is one of the phonological rules active in Old Norse, we shall come back to it at great length in the near future.

In addition to this rule for narrowing the double consonant, Spurkland summarizes the realization of the dental suffix in four main rules.

• If the root ends in a voiced consonant, i.e. /v/ (the spelling is always <f>), /r/, /g/, /l/, /m/, /n/, the dental suffix realizes as /ð/. He gives the following examples.

kraf-ð-i, bar-ð-i, fylg-ð-i, dǿm-ð-i, val-ð-i, van-ð-i

Let's take a closer look at fylg-ð-i (illustrating the telja-class) and døm-ð-i, for example.

fylgja - (fylgir) - fylgði - (fylgði) - fylgt

dǿma - (dǿmir) - dǿmði - (dǿmði) - dǿmt

Our original dental suffix, the Proto-Germanic heir, persists in this case.

• If the root ends in a voiceless consonant, the dental suffix becomes the voiceless plosive /t/. Spurkland illustrates with vak-t-i, flut-t-i, merk-t-i.

This voiceless root ending consonant can be /t/ itself as in vakta or flytja.

vak-t-i knows the evoked double consonant reduction

vakt-t-i→vak-t-i

which does not occur in flut-t-i because the geminate is not preceded by a consonant.

• If the root ends in /ð/, the dental suffix realizes as /d/ and a regressive assimilation of the root /ð/ into /d/ occurs. In short, the <ðð> appearing in conjugation turns into <dd>. The phonological reason is simple: there is no such thing as a long /ð/, i.e. a /ðð/.

We had sampled, for example,

leiða - (leiðir) - leiddi - (leiddi) - leitt
hræða - (hræðir) - hræddi - (hræddi) - hrætt

where the past goes through

leið-ð-i→leið-d-i→leid-d-i
hræð-ð-i→hræd-ð-i→hræd-d-i

If the root ending /ð/ is preceded by a consonant, naturally, the shrinking further occurs. For example, virð-a does

virð-ð-i→virð-d-i (no double-ð)→vird-d-i (regressive assimilation)→vir-d-i (reduction)

• If the root ends in /d/ preceded by another consonant, the dental suffix realizes as (is assimilated into) /d/. Since the geminate <dd> is preceded by a consonant, it is reduced according to the rule evoked.

Spurkland cites here

senda - (sendir) - sendi - (sendi) - sent

The contamination-reduction go as follows:

send-ð-i → send-d-i →sen-d-i

Nevertheless, those rules seem not to account for at least two of our earlier finds

stefna - (stefnir) - stefndi - (stefndi) - stefnt
sóma - (sómir) - sómdi - (sómdi) - sómt

just as Haugen's rules do not foresee val-ð-i or van-ð-i. Spurkland preempts the objection:

Nå følges ikke disse reglene helt sklavisk. Som så ofte ellers er det tale om en tendens og en tendens i utvikling.

i.e., “Now, these rules are not followed slavishly. As is so often the case, this is a trend and an evolving trend.” And spurkland cites a few more exceptions (to his own rules).

For det første kan vi finne ustemt apikalsuffiks der vi aller minst skulle vente å finne det: mæla (= snakke) - mælti, ræna (= rane) - rænti. Dessuten kan man ofte finne en veksling mellom plosiv og frikativ (dvs. [d] og [ð] etter samme konsonant: fella - (*felldi →) feldi, telja - talði; kenna (*kenndi →) kendi, venja - vandi.

In short, Spurkland's rules rob Peter to pay Paul, but linguists are well aware of the limits of any attempt at rules: languages are in constant motion. Linguistics attempts - and this is their scientific ideal - to systematize observed human facts. The question of what linguistics is in the machine age is an interesting one, and one we shall shortly enjoy wrestling with.

Finally, as some excellent Old Norse reference grammars, quoted at length in the foregoing, are by Norwegian and Nordic scholars, we hope to have convinced you a little further to convert to the Norwegian language.

References

Ringe, D. (2017). From Proto-Indo-European to Proto-Germanic (2nd ed.). Oxford University Press.

Spurkland, T. (1989). Innføring i norrønt språk. Universitetsforlaget.

But what about authorship in the Middle Ages, and more specifically, for our purposes, in the Nordic Middle Ages? Heimskringla, John Megaard reminds us, has an exceptional status: indeed, it is inseparable from the name of Snorri Sturluson, while, otherwise, “the great saga literature has almost always been handed down without an author's name, and in cases where an author's name is found, it often belongs to a person whom we know little about.” (Megaard, 2017, p.18-19) John Megaard, philologist and specialist in Icelandic sagas, gives his remarkable essay tracking down these medieval authors a paradoxical title: Hvem skapte Heimskringla? i.e., Who created Heimskringla? It is the basics of his method that interest us today, but let's return for a moment to the question of authorship in Snorri's time.

Subscribe now

We need to think back to the conditions in which the written word was produced and disseminated before Gutenberg. The term “author” and, in today's society, the intellectual property rights associated with it, are only valid when, firstly, the works are clearly delineated and, secondly, they can be published, i.e. made amply accessible to the public.

The manual work of writing was tedious, akin to the trade, that of a carpenter or goldsmith, and hence the book itself a precious object. The works of Snorri's time are handed down by a number of transcribers likely to leave their mark in different ways: at their discretion, to shorten, or to append or to replace words or expressions in the interests of clarity. So much so that some scholars, most notably Iceland's Sigurður Nordal, argue that the idea should be abandoned of ascribing sagas texts to authors on language grounds. This opinion has kept the debate on the authorship of sagas at bay for some time.

Beyond transcription, it is also editing, particularly by historians, that blurs the boundaries of works. Since antiquity, it has been common for them to extend a work that already had several contributors. Here, The saga of Sverre is highly representative. Primarily edited by Abbot Karl Jonsson, its original story seems to have been written by two residents of the same cloister, and its prologue mentions two other contributors to the publication. (Megaard, 2017, p.24-25)

The question of publicity, for its part, ties in with that of the works' destination. Here, Holzknecht’s work on Literary Patronage in the Middle Ages is illuminating. He describes the progressive emergence of authorship, already observed in societies before Biblical times. “There had been first of all a communal period of cooperative effort in which individual ownership gradually developed as the literary individual emerged from the throng.” Recognition of a literary profession confronts authors with the prevailing economic conditions, while recognition of literary distinction awakens them to their market potential. For remuneration, “it is to individuals that the poet must look”. Then, it is “only when the author perceives that his poetry is a ware which may possibly be sold, and decides […] to court an individual with it, that the real sense of literary property develops and patronage as a source of compensation comes into being”. This emergence movement is not unidirectional across the ages. In the troublous times of the early Middle Ages, “the high individualism of Empire slipped back into an almost communal condition”, from which a slow, “gradual regrowth of the conditions of classical times" operates. The latter was catalyzed by the rise of the highly modern “ideas of fame, of glory, of individual achievement, literary property and individual worth”, which characterize the Renaissance.

Patronage, ultimately, “is the employment”, regardless of its motives, “of favor, protection and influential support to advance the interests of art”, particularly by subsidizing authors. It almost always implies, as the name suggests, “something of the relation of superior to inferior which existed between a powerful Roman patron and his clients”. A highly inegalitarian distribution of wealth makes it a necessity; "patronage, Holzknecht tells us, flourished best in a highly aristocratic form of society", and "in this respect, medieval society was ideal". Although “sporadic and uncertain”, and of a magnitude incomparable with that of the Renaissance, medieval patronage was the province of kings and emperors, driven often by a fondness for literature: mécénat became a fashion. (Holzknecht, 1923, p.2-3; p.55-56)

Authors of the Nordic Middle Ages, Megaard tells us, write for “very specific addressees”: their works of art are “in the recipient's honor”. This is clearly demonstrated in poetry: the skald 's undying attentiveness to the king or patron from whom he can expect a reward, expresses itself in panegyric of the latter or detraction of his enemies. Such is the case of the poet Snorri, in Håttatal, a tribute poem to Håkon Håkonsson and Duke Skule. Another saga, Sturla Tordsson's about Håkon Håkonsson, serves as a source and corroboration: we learn that Duke Skule commissioned Snorri to compose a poem disparaging a court man, Gaut Jonsson, alleged to have ruined Skule's friendship with the king. (Megaard, 2017, p.38-39)

Megaard's precise aim is to demonstrate that the prose of sagas, like the poetry of skalds, exhibits what he calls tendencies (tendens): evidence, running through the texts, of some kind of partisanship. These tendencies then serve as ideal tools for the philologist seeking to establish authorship.

Share

John Megaard’s method is precisely interested in the genesis of works from ancient Icelandic times: both history and literature. It is applied, in his book, to the unraveling of the mystery surrounding Heimskringla, a work yet commonly referred to, as if its author were unequivocal, as Snorri's Kings' Sagas (Snorres kongesagaer). The scholar Lars Lönnroth seems to have been the first, in 1965, to cast doubt on Snorri's authorship of both the Edda and Heimskringla. He demonstrates that the authors of the sagas, important intellectuals, were commissioned by heads of state, and reexamines Snorri's role as either author or compiler. (Lönnroth, 1965)

Several troubling elements around Snorri’s Edda fuel these suspicions. All the manuscripts differ from one another, with only Codex Trajectinus not adding to the three parts evoked in the introduction. This introduction, in its Uppsala version, is both precious and ambiguous. It says

Bók þessi heitir Edda. Hana hefir saman setta Snorri Sturlu sonr eftir þeim hætti, sem hér er skipat. Er fyrst frá ásum ok Ymi, þar næst Skáldskaparmál ok heiti margra hluta, síðast Háttatal, er Snorri hefir ort um Hákon konung ok Skúla hertoga.

“This book is called Edda. Snorri Sturluson has compiled it in the manner arranged here. First it tells of the Æsir and Ymir, then Skáldskaparmál (the language of poetry) and names of many things, finally Háttatal which Snorri has composed about King Hákon and Duke Skúli.”

Snorri wrote Håttatal, and we can draw no conclusions about his authorship of the other two parts. How should “has compiled” be interpreted? Another source, the medieval Icelandic manuscript Oddaverjaannáll, reiterates: “Han satte sammen Edda og mange andre lærde bøket [og] islandske sagaer.” Megaard explains:

On the one hand, the juxtaposition of "Edda" and "many learned books" could give reason to interpret "put together" as "translate", "organize", "collect". But this is less appropriate for “Icelandic sagas”, where the meaning “to author” can appear quite natural, at least by our standards.

Manuscript variability is perhaps what prompts the author of this preamble to specify his version faithful to Snorri's original: "Snorri Sturluson has compiled it in the manner arranged here". (Megaard, 2017, p.250-251)

Subscribe now

As the authorship of works, even those most assertively attributed by philological tradition, comes into question, investigations begin. John Megaard’s method postulates that tendencies, i.e. marks of partisanship, pervade the texts.

For an example of the tendencies at work in the Heimskringla, we can turn to two consecutive sagas: The saga of King Magnús Barefoot, followed by that of his three sons, The Saga of Sigurðr the Jerusalem-farer, Eysteinn and Óláfr. The study of those tendencies points Megaard towards an attribution to Are Frode, the better-attested author of Íslendingabók, which we have perused many times.

Here are his conclusions.

Are [...] synes å ta parti for Magnus Berrføtts motstandere i kampen om tronen, samtidig som de har sympati for de han slåss mot på hærtogene utenfor Norges grenser.

“Are [whom Megaard assumes have written or contributed to the Saga] [...] seems to side with Magnús Barefoot's opponents in the battle for the throne, while at the same time having sympathy for those he fights against on army campaigns outside Norway's borders.”

Subsequently, Sigurðr the Jerusalem-farer and Eysteinn, Magnús Barefoot's successors to the throne, are described by the presumed author as pursuing widely diverging foreign policies in Northern and Southern Norway, over which they respectively reigned. The relationship to Sweden is particularly telling. Eysteinn strengthens ties with Magnús Barefoot's former opponent through marriage. Sigurðr allies himself with the Danish crown, and immediately after his brother's death, calls for an offensive against Sweden.

What follows is particularly illuminating about the John Megaard method. On Eysteinn's death, a new royal pretender (kongsemne) is put forward to succeed him. As the saga depicts it, Sigurðr has no choice but to give in:

Fortellingen […] i Heimskringla tyder på at det nye kongsemnet har hatt sterke støttespillere, siden Sigurd Jorsalfar måtte gi etter og godkjenne slekskapet.

“The story [...] in Heimskringla suggests that the new royal pretender had strong supporters, since Sigurd Jorsalfar had to give in and approve the kinship.”

The royal pretender, Harald Gille, would reign over half of Norway from the time of Sigurd's death, before definitively defeating his rival Magnus Blinde four and a half years later. Our author, Megaard says, presents the events as a victory for Eysteinn's party, allied with the Swedes, over that of Sigurd Jorsalfar, allied with the Danish crown. The tendencies are reflected in the choice of semantics, whether laudatory or disparaging, but above all in the choice of an angle of view (synsvinkel). Are Frode's successors, he concludes, will present the history of kings differently. (Megaard, 2017, p.213-215)

However, the collection of clues is not limited to bias. As we have seen, Megaard also summons the manuscripts. Another crucial aspect of his method comes into play here: to each of the five authors screened as potential contributors to Heimskringla, at least one text can be attributed with certainty, providing a solid foundation for meaningful analysis. So, the work most solidly attached to Are Frode is his Íslendingabók - it opens, as we read earlier, with an account of its sources, the author's ancestors, who coincide well with Are's. We have information about the identity of the pretenders to authorship Isleiv, Sæmund, and Are. While information about Eirik Oddsson is scarce, his authorship of Hryggjarstykki is well documented. Certainty about both reference works and subsequent attributions emerges indeed from confronting documents: manuscripts, secondary literature, author information, the intersections of history and biography.

Finally, Megaard does not disregard literary evidence, finding striking similarities between seemingly distant works. Among these are stylistic features - in Are Frode's case, a distinctive way of introducing word synonyms. Or the way the text is interrupted to showcase the verses of the skalds, who we have seen clearly composed to the glory of kings.

Share Old Norse Linguist

Ultimately, John Megaard's investigations serve the purpose of “dating Old Norse prose literature” (Megaard, 2017, p.42)

[…] ut fra et ønske om å bidra til dateringen av den norrøne prosalitteraturen.

and generally shedding light on the texts’ genesis, for two ends. First, to examine the value of the sagas as historical sources, regarding the history of kings, the portrait of society both from the time of the narrated history and from the time of writing. Secondly, for the sake of philology and to reconstruct the genre's history.

We said above

The term “author” and, in today's society, the intellectual property rights associated with it, are only valid when, firstly, the works are clearly delineated and, secondly, they can be published, i.e. made amply accessible to the public.

This delimitation is blurred by our increasingly inevitable and inextricable intermingling at work with intelligent machines; access to publication, meanwhile, is threatened by its excess. Overproduction by ever more educated masses, instantaneous, ubiquitous and fluid internet distribution, asymptotically induce perfect deafness: the paroxysmal possibility of addressing absent subscribers. The man of the future will be much less sensitive to the concept of authorship than the John Megaard of early 21st century. He will be like the troubadour or minstrel of ancient communal times, wholeheartedly dropping a handcrafted trickle into the ocean of machine words.

Subscribe now

References

Megaard, J. (2017). Hvem skapte Heimskringla?: forfatterne bak Snorres kongesagaer. Solum Bokvennen.

Holzknecht, K. J. (1923). Literary Patronage in the Middle Ages. University of Pennsylvania.

Lönnroth, L. (1965). European Sources of Icelandic Saga-Writing: An Essay Based on Previous Studies. Thule

Just compare. We dive into the dream of the mad king Sigurð, in The Saga of Sigurd the Crusader, Eystein and Olaf.

If you are a native Norwegian speaker, or already a convert, you can perhaps divine the saga's meaning simply by reading the Old Norse. Nevertheless, you will be delighted to read more about Sigurd, and will certainly learn to examine your language in more detail: we are often unaware of the inner workings of our own language. If not, we hope to demonstrate sufficient proximity between Old Norse and modern Norwegian to convert you to the entertaining apprenticeship of a northern language whose music is incredibly charming.

King Sigurd becomes taciturn.

Draumr Sigurðar konungs Jórsalafara.

Þat er sagt, at Sigurðr konungr tók úgleði mikla, ok mátti lítt njóta tals hans, ok sat hann skömmum við drikkju.

Þat þótti ráðsmönnum þungt, ok vinum hans ok hirðinni, ok biðja Eystein konung leggja til nökkur ráð, at hann féngi vitat, hverju gegndi; féngu menn nú enga orskurði mála sinna, er hann sóttu at. Eysteinn konungr svarar svá: Vant er at rœða ok leita eptir við konung.

Get acquainted (you're not Norwegian) with the story.

The Dream of King Sigurd the Crusader.

It is said that King Sigurd fell into a great melancholy, and people could hardly engage him in conversation, and he sat but briefly at drink.

This weighed heavily on his counselors, friends, and retinue, and they asked King Eystein to devise some plan so that he might learn what was the matter; for men now received no decisions in their cases when they sought his counsel. King Eystein answered thus: It is difficult to speak with and inquire of the king.

Subscribe now

Let's take a closer look at an Norwegian equivalent to the first sentence:

Þat er sagt, at Sigurðr konungr tók úgleði mikla, ok mátti lítt njóta tals hans, ok sat hann skömmum við drikkju.

First of all, an almost word-for-word translation, with loose syntax, but completely comprehensible to a native Norwegian.

Det er sagt, at Sigurd konge tok uglede mye, og kunne lite nyte tale hans, og satt han kort ved drikke.

We highlight below obvious similarities.

Þat er sagt, at Sigurðr konungr tók úgleði mikla, ok mátti lítt njóta tals hans, ok sat hann skömmum við drikkju.

Det er sagt, at Sigurd konge tok uglede mye, og kunne lite nyte tale hans, og satt han kort ved drikke.

Almost everything. Let's try to convey the feel of this approximation for the English-reader.

It is said, that Sigurd king took sadness (*un-gladness) much, and could little use talk his, and sat he briefly at drink.

Let's highlight some immediate similarities, this time from Old Norse to our broken English.

Þat er sagt, at Sigurðr konungr tók úgleði mikla, ok mátti lítt njóta tals hans, ok sat hann skömmum við drikkju.

It is said, that Sigurd king took sadness (*un-gladness) much, and could little use talk his, and sat he briefly at drink.

Some of the word-for-word English is very recognizable. But even in pieces that coincide rather well, English is clearly further from Old Norse than Norwegian is:

Þat er sagt - Det er sagt - It is said

at Sigurðr konungr tók - at Sigurd konge tok - that Sigurd king took

ok (mátti) lítt njóta tals hans - og (kunne) lite nyte tale hans - and could little use talk his

sat hann (skömmum) við drikkju - satt han kort ved drikke - sat he briefly at drink

The present meditation is all about comparative linguistics, and dreams in particular of clarifying the notion of distance between languages. As we said earlier

Galois dreams of alchemizing a precise, very mathematical notion of distance between languages. A machine which, given the universal corpora of any two languages, engineers their objective score of similarities, in all respects, according to a secret and exact formula.

Even if our great objective still requires a little work, we have no difficulty in observing the variations in proximity highlighted above.

Let's look ahead to smoother, more syntax-accurate translations in both languages. In a first version, we re-establish the Norwegian syntax, and in particular the word order that was previously broken.

Þat er sagt, at Sigurðr konungr tók úgleði mikla, ok mátti lítt njóta tals hans, ok sat hann skömmum við drikkju.

Det er sagt at kong Sigurd tok stor uglede, og kunne litt nyte talen hans, og (han) satt kort ved drikken.

Although not quite modern, we have both correctness in Norwegian, and closeness to Old Norse. (Among the awkwardnesses, tok stor uglede, literally took big enjoyment, and kunne litt nyte talen hans, literally, could (just) a little use his speech (or talk).)

We can propose a more literary and elegant translation,

Þat er sagt, at Sigurðr konungr tók úgleði mikla, ok mátti lítt njóta tals hans, ok sat hann skömmum við drikkju.

Det sies at kong Sigurd ble grepet av stor tungsinn, og man kunne knapt føre en samtale med ham, og han satt kort tid ved drikkebordet.

This is the fluid English translation we were proposing.

Þat er sagt, at Sigurðr konungr tók úgleði mikla, ok mátti lítt njóta tals hans, ok sat hann skömmum við drikkju.

It is said that King Sigurd fell into a great melancholy, and people could hardly engage him in conversation, and he sat but briefly at drink.

Subscribe now

Here again, even the most legible pieces of Old Norse into English have a more immediate Norwegian equivalent.

Þat er sagt, at Sigurðr konungr - Det sies at kong Sigurd - It is said that King Sigurd

ok sat hann - og han satt - and he sat

við drikkju - ved drikkebordet - at drink

Let's take a look at the rest of King Sigurð's story.

Þat þótti ráðsmönnum þungt, ok vinum hans ok hirðinni, ok biðja Eystein konung leggja til nökkur ráð, at hann féngi vitat, hverju gegndi; féngu menn nú enga orskurði mála sinna, er hann sóttu at. Eysteinn konungr svarar svá: Vant er at rœða ok leita eptir við konung.

Here again is a suggestion for a word-for-word translation, with questionable syntax, but very readable. We highlight clear similarities.

Þat þótti ráðsmönnum þungt, ok vinum hans ok hirðinni, ok biðja Eystein konung leggja til nökkur ráð, at hann féngi vitat, hverju gegndi; féngu menn nú enga orskurði mála sinna, er hann sóttu at. Eysteinn konungr svarar svá: Vant er at rœða ok leita eptir við konung.

Dette syntes rådsmennene tungt, og vennene hans og hirden, og ba kong Eystein legge til noen råd, at han fikk vite hva det gjaldt; fikk menn nå ingen avgjørelser av sakene (målene) sine, som han søkte etter. Kong Eystein svarer så: Vanskelig er det å tale og leter etter ved kongen.

A word about mála, for example. This is the genitive plural of the substantive mál, meaning language, word, but in this case matter, affair, or perhaps even more precisely, cause, in a legal context. In Norwegian Bokmål, there are two etymologies for mål. The descendant of mál admits language as the main meaning (as in bokmål), but the meaning cause persists, notably in the compound nouns giftermål (marriage), klagemål (complaint), spørsmål (question, inquiry), søksmål (lawsuit), veddemål (wager). The Norwegian speaker, by constantly delving into the depths of his language, can easily intuit the meaning of Old Norse mála sinna.

Share

Finally, we can offer two fluid, modern translations, and perusing them will convince the reader of this historical truth: Norwegian is much closer to Old Norse than English.

Dette tynget rådsmennene, vennene hans og hirden, og de ba kong Eystein komme med noen råd, slik at han kunne finne ut hva som feilte; for nå fikk folk ingen avgjørelser i sakene sine når de oppsøkte ham. Kong Eystein svarer slik: Det er vanskelig å snakke med og utspørre kongen.

This weighed heavily on his counselors, friends, and retinue, and they asked King Eystein to devise some plan so that he might learn what was the matter; for men now received no decisions in their cases when they sought his counsel. King Eystein answered thus: It is difficult to speak with and inquire of the king.

The answer to the initial question, Hvorfor lære norsk, i.e. why learn Norwegian, takes shape: It's much simpler to understand Old Norse if you know good Norwegian.

Subscribe now

References

The Saga of Sigurd the Crusader, Eystein and Olaf. In Heimskringla III. In B. Aðalbjarnarson (Ed.). (1941-51). Íslenzk fornrit (Vol. 28). Hið íslenzka fornritafélag.

What about weak verbs? Let's look briefly at the German for to cook or to love.

lieben - liebt - liebte - liebten - geliebt (to love)
kochen - kocht - kochte - kochten - gekocht (to cook)

Those are German weak verbs, and as with English weak verbs, the stem vowel is not altered by conjugation. The ending -te/-ten is added to the present tense stem to form the preterite, -t to form the past participle. As in English, German weak verbs form their preterite and past participle by means of a dental suffix, that is to say, some /t/ or /d/ sound or similar: this, anew, dates back to Proto-Germanic times.

We shall come back to this Proto-Germanic origin in a moment, but for now, let's take a look at Old Norse's realization of this dental suffix, we said, some /t/ or /d/ sound or similar.

Subscribe now

For example, let us examine a passage from Heimskringla, The Saga of Sigurd the Crusader, Eystein and Olaf, where Chapter 23 appears rich in weak verbs in their preterite form or past participle.

Ólafr konungr tók sótt, þá er hann leiddi til bana; ok er hann jarðaðr at Kristskirkju í Niðarósi; ok var hann hit mesta harmaðr. Síðan réðu þeir tveir konungar landi, Eysteinn ok Sigurðr, en áðr hǫfðu þeir þrír brœðr verit konungar 12 vetr, 5 síðan er Sigurðr kom til lands, en 7 áðr. Ólafr konungr var 17 vetra, er hann andaðist, en þat var 9 Kalendas Januarii.

The dental suffix realizes as <d> in leiddi, and <ð> in jarðaðr, harmaðr, hǫfðu, andaðist. Let's collect other examples from the same saga.

… en eptir dvaldist mikill fjöldi Norðmanna …

… Þá mælti Sigurðr konungr …

… ok tók þá bók hina dýru, er hann hafði haft í land …

… ok var þessi leiðangr kallaðr Kalmarna leiðangr …

… ok lagði hann þar mikit fé til …

… ok þakkaði várum dróttni feginsamliga …

… Þrimr vetrum síðar en krosskirkja var vígð …

… ok hræddist bæði sult ok píslir …

… Þá stefndi Sigurðr konungr máli þessu til Arnarnesþings …

… en þeim öllum viltist sýnin …

… hann gipti eina dóttur sína Heinreki keisara …

… heimti hann Ívarr til máls við sik …

… en þann veg skipti þá til, sem …

… var þat sagt, at …

… ok gerðist þá ilt til matar …

… Þá er Sigurðr konungr sigldi fyrir Spán …

… meðan hann lifði …

… sem honum sómdi at veita ríkum mǫnnum …

We are now witness to three realizations of the suffix dental, which anew forms in the highlighted examples either the preterite or the past participle. Namely, we find a dental suffix realized as

<d> in dvaldist, hræddist, stefndi, sigldi, sómdi,

<t> in mælti, haft, viltist, gipti, heimti, skipti, sagt,

<ð> in hafði, kallaðr, lagði, þakkaði, vígð, gerðist, lifði

Can we draw a model from these observations, i.e. can we predict from the infinitive form whether the dental suffix of the preterite (and past participle) will be <d>, <t> or <ð>?

A twin question from a historical perspective: what historical factors determine which form of the dental suffix appeared in Old Norse weak verbs? Note the important difference. Both questions seek to explain the realization of the dental suffix as <d>, <t> or <ð>. However, the first is synchronic, i.e. it looks for explanations within the fixed temporal framework that corresponds to Old Norse. The second is diachronic, and examines the evolution of languages from Proto-Germanic to Old Norse, via Proto-Norse and their close relatives. Of course, knowing how the observed phenomenon derives historically paves the way for a synchronous explanation. (This duality of synchronic and diachronic research is captivating, and we shall devote future reflections to it.)

But first, back to our synchronic observation. Typically, the phonological environment influences orthography. Here, we notice promptly that, in all our examples above, the dental suffix is realized as <ð> after a vowel: kallaðr, þakkaði, jarðaðr, harmaðr, hǫfðu, andaðist. With consonants, the findings are less obvious. We can indeed collect, regarding the preterite forms,

<l> (dvaldist, sigldi), <n> (stefndi), <d> (hræddist, leiddi), <m> (sómdi) before <d>

<l> (mælti, viltist), <p> (gipti, skipti), <m> (heimti) before <t>

<f> (hafði, lifði), <g> (lagði), <r> (gerðist) before <ð>

and regarding the past participles

<f> (haft), <g> (sagt) before <t>

<g> (vígð) before <ð>

Ideally, we would like to conclude something like: the realization of the dental suffix after a consonant depends on the "consonant type". And for that, we need to say a brief word at least about Old Norse consonants.

Share Old Norse Linguist

Phonetics distinguishes consonants by certain phonetic characteristics, the most important of which are

• Manner of articulation, e.g. stops, fricatives and nasals;

• Place of articulation, i.e. where in the vocal tract the consonant is obstructed, and which speech organs are involved, for instance, a bilabial (both lips), alveolar (tongue against gum ridge) or velar (tongue against soft palate) articulation;

• Phonation, i.e. whether the vocal cords vibrate during articulation (“voiced”) or not (“voiceless”).

Along the lines of this typology, Old Norse consonants are arranged as follows.

(We follow here closely Wikipedia, Old Norse, Consonants, and the description is similar in Haugen.) (Haugen, 2015, p. 21)

A word about notations. Up to now, we have noted between brackets <>, the grapheme, i.e. the letter as written. In the preceding phonology table, these are not graphemes, but phonemes, sounds, usually noted between slashes //. The correspondence is not totally bijective. For example, the letter <g> can be realized by the phonemes /g/, /ŋ/ or /ɣ/, depending on the phonological context. In the table, the phonemes in (parentheses) have no corresponding grapheme, this is also the case for /θ/ and /ð/, both phonemes being rendered as <θ> in standardized spelling. Let's note in passing that when the grapheme-phoneme correspondence is unique, the /sound/ and the <letter> are unambiguously equivalent. Back to our consonants.

Subscribe now

We won't describe the whole system today, just take a closer look at the consonants we have collected.

Before the dental suffix /d/ we find /l/, /n/, /m/, all of which are voiced (/l/ can be both).

Before /t/, we find the voiceless consonants /r/, /l/, /p/. (We can note an obvious exception heimti, /m/ being voiced.)

And all phonetic realizations of <g> as well as /r/, found before /ð/, are voiced. <f> in hafði or lifði is realized as /v/ which is voiced too.

These are great finds! In short, the dental suffix, in the formation of the past tense and past participle of weak verbs, is realized in

• /t/ (voiceless) after a voiceless consonant;

• /d/ (voiced) after /m/, /n/ or /l/, three consonants that are a little special to Old Norse, as they classify, just like vowels, as sonorants;

• /ð/ (voiced) after a vowel (this linking vowel a of the first class of weak verbs) or after a voiced consonant.

  

The voiced or voiceless character of the dental suffix seems to arise through contamination. A kind of proximity rule emerges, which we could now try to trace diachronically.

Let's return to the phenomenon's proto-Germanic ancestry, which explains its dissemination across the Germanic spectrum. That's what we said in our epic of Old Norse weak and strong verbs.

What characterizes all weak verbs, Ringe tells us, is the common formation of the past participle, of a default finite past tense (all tenses, modes and persons), with the exception of a distinct singular indicative past tense. More specifically, the past participle of weak verbs always has a stem vowel *-a- or *-ō-, and is formed with the suffix *-da- (with a few exceptions). The past tense suffix begins with the dental obstruent *-d-, with the suffix and endings following a well-defined pattern. (Ringe, 2017, p.280) (To level out the terminology, a conjugated verb is formed by a stem, a suffix (here, the past tense suffix begins with a *-d-) and an ending typical of the mode, tense and person.)

First of all, the dental suffix is already present in Proto-Germanic, that most recent ancestor - reconstructed by linguistics - common to all subsequent Germanic languages. Although classically spelled <d>, it seems to be have been pronounced as the voiced fricative /ð/. It would have hardened to the plosive /d/ in the West-Germanic languages while remaining a fricative in the North-Germanic branch, including Old Norse.

Here is anew (the relevant extract from) our phonetic table.

The Icelandic case corroborates our hypothesis of contamination by context. Gobiirsch, although presenting a later Icelandic phonological phenomenon in his article, tells us (Gobiirsch, 2021)

While NWGm b, d, g are phonologically voiceless, they may be partially or fully voiced medially in a voiced envi­ronment.

In Old Icelandic, obstruents get voiced and devoiced based on their phonological environment.

There may also be voicing of the NWGm fricatives medi­ally in a voiced environment. As in all northern North-West Germanic (Old Norse, Old English, and Old Saxon), there was only one series of fricatives with complementary distribution of voiced and voiceless variants in Old Icelandic. The loss of voicing contrast in fricatives present in Common Germanic was due to the final devoicing of Gmc β, ð, γ and the voicing of Gmc f, θ, x medially in a voiced environment (cf. Goblirsch 1999b).

Old Icelandic, like Old Norse, devoices the dental suffix /d/ in /t/ after a voiceless consonant, and in some other cases (heimti, found in the same way in modern Icelandic, may be one of those).

In fact, the phonological reciprocal influence of the suffix dental on its surroundings is plausible, if we read Ringe carefully:

Immediately before *t, all labial consonants were replaced by *f and all dorsal consonants by *h. This can be seen in inflectional forms such as past singular *gaft 'you gave' (compare Gothic and Old Norse gaft, from *gebaną 'to give') and present singular *maht 'you can' (compare Old English meaht, Old High German maht, from *maganą 'to be able').

He speaks here of Proto-Germanic transformation. (Ringe, 2017, p.247) A number of regular transformations transform language over the generations, from Proto-Indo-European to Proto-Germanic, then on towards modern languages. One of the most important of these laws is Grimm's Law, named after the famous 17th-century philologist of Germanic languages and writer. Grimm’s law, tells us Ringe, can be seen as a “chain shift” where the component sound change together. It consists, thus, of several sound changes, that are active in the transition from Proto-Indo-European to Proto-Germanic. By the most significant of them,

PIE voiceless stops became PGmc fricatives, provided that they were not immediately preceded by another obstruent (usually *s, but sometimes another stop).

(Ringe, 2017, p.113-122)

This might well justify our dental suffix rendered as the fricative /ð/, at the Proto-Germanic level. Ringe doesn't specify this precise case, and exact genealogy is not possible in such a short space. But we do appreciate the overall historical movement.

We shall be coming back to Grimm's law, its successor (in the sense that the variations it describes occur, historically, after Grimm's chain shift), Verner's law, and, naturally, the morphology and phonology of Old Norse.

References

The Saga of Sigurd the Crusader, Eystein and Olaf. In Heimskringla III. In B. Aðalbjarnarson (Ed.). (1941-51). Íslenzk fornrit (Vol. 28). Hið íslenzka fornritafélag.

Wikipedia, Old Norse, Consonants, https://en.wikipedia.org/wiki/Old_Norse#Consonants

Haugen, O. E. (2015). Norrøn grammatikk i hovuddrag. Novus.

Ringe, D. (2017). From Proto-Indo-European to Proto-Germanic (2nd ed.). Oxford University Press.

Goblirsch, K. G. (2001). The Icelandic consonant shift in its Germanic context. Arkiv för nordisk filologi, 116, 117-133.

We have just learned that Are Frode, who authored Íslendingabók, got his story from very reliable sources, his ancestors, and that the colonization of Iceland took place “from Norway in the days of Harald Fairhair”. Let's carry on. We are still in chapter one, Frá Íslands byggð, that is, About the Settlement of Iceland.

Subscribe now

A hearty piece of Old Norse for today.

Ingólfr hét maðr nórrænn, er sannliga er sagt, at færi fyrst þaðan til Íslands, þá er Haraldr inn hárfagri var sextán vetra gamall, en í annat sinn fám vetrum síðar. Hann byggði suðr í Reykjarvík. Þar er Ingólfshöfði kallaðr fyr austan Minþakseyri, sem hann kom fyrst á land, en þar Ingólfsfell fyr vestan Ölfossá, er hann lagði sína eigu á síðan.

Í þann tíð var Ísland viði vaxit á milli fjalls ok fjöru. Þá váru hér menn kristnir, þeir er Norðmenn kalla Papa, en þeir fóru síðan á braut, af því at þeir vildu eigi vera hér við heiðna menn, ok létu eftir bækr írskar ok bjöllur ok bagla. Af því mátti skilja, at þeir váru menn írskir.

Let's take a look at the English translation. It is perfectly legitimate to go through it right now! Extensive reading consists precisely in reading the original text alongside its translation into a well-mastered language, and practicing with the utmost concentration to find, word group by word group, the correspondences between the two versions.

There was a Norwegian man named Ingolf, who it is truly said went first from there to Iceland when Harald Fair-hair was sixteen winters old, and a second time a few winters later. He settled south in Reykjavik. There is a place called Ingolf's Headland east of Minthak's Eyri, where he first came ashore, and there is Ingolf's Mountain west of Olfus River, where he later claimed his property.

At that time, Iceland was wooded between mountain and shore. There were Christian men here, whom the Northmen called Papar, but they later departed because they did not want to stay here with heathens, and they left behind Irish books, bells, and croziers. From this it could be understood that they were Irish men.

The pioneer of Icelandic colonization was a Norwegian named Ingolf, and the first sentence is simple. We highlight the main clause in blue.

The English here is not very literal. Closer to the text, we could translate this as Ingolf was named the Norse man, who… The formula is very common in the sagas. Here, the sentence's grammatical subject maðr nórrænn is positioned after the verb, and the predicate nominative Ingólfr - how maðr nórrænn is named - is in first position. Old Norse, as already mentioned, has a very rich inflectional system, so the subject-verb-object order, though very frequent, is not obligatory. In short, since grammatical cases - subject or object, for example - are clearly identifiable by their form - their suffix, their ending - Old Norse practitioners can take a few liberties with word order: the addressee can reverse-engineer the function by following the declension table.

Share

What follows is a long relative proposition, of antecedent maðr nórrænn, and comprising a main (parenthetical) passive construction, er sannliga er sagt and a subordinate which itself comprises two time adjuncts connected by the coordinating conjunction en. We will become masters in the art of identifying this type of structure, just as we will become strong in conjunctions, relative particles and, in short, in connecting words.

First, the first occurrence of er, which could very well be the verb vera, to be, conjugated in the present indicative in the 3rd person, but which, in context, is the relative particle, a versatile subordination marker translatable as who or one of his close relatives. How do we know? Well, this clause already has a conjugated verb: precisely the second er, and a perfectum participle , sagt, is said.the second er! Not on its own, but accompanied by the perfectum participle sagt, from segja (to say) to form the complete finite form in the passiv er sagt, is said.

At functions as a complement (a subordinating conjunction) in Old Norse, introducing a complement clause, in this case a reported speech complement. Note that in English, this subordinating conjunction is not expressed. But we could rephrase a little, to stay closer to the Old Norse.

We are closer to the Old Norse because, firstly, the relative clause maintains the present passive verb, is said, and secondly, that appears, which directly mirrors the subordinating conjunction at. Our Old Norse relative particle, er, is invariable, so how can it distinctly translate the declensions of English who: whom or whose? Well, these variations can be rendered by association with a demonstrative pronoun - we shall have ample chance to illustrate this - but also simply by the structure of the relative proposition, which is the case here. But this relative particle can also accompany an adverb, as with þá er, and the whole functions as a modern subordinating conjunction, when: literally, then as. Speaking Norwegian, as we shall often say, can clear up a lot of ambiguities. The particle er can be translated fairly systematically into som, which can either be the Norwegian relative pronoun or mean as (in the comparison of equals, as big or as short as). þá translates transparently into the Norwegian da and the resulting *da som is perfectly comprehensible: the Norwegian would say da on its own. Finally, you need to recognize en, but (or and), immediately to best cut sentences into pieces.

The preceding syntactic analysis clearly illuminates the parallel framework of Old Norse and English. But we can work a bit more to understand this passive construction, er sannliga er sagt, which we have surreptitiously called parenthetical. The latter does not change the course of events, but only informs about the reliability of the sources, that this story is reported, and true. What is really the action of which Ingólfr is the agent? He went to Iceland. We shall rephrase a little more, as follows.

We are now highlighting our functional relative clause, which coincides with the syntactic relative of our new version. The inserted clause about the narrative can certainly be dispensed with altogether, all things being equal: a parenthesis, indeed.

Share Old Norse Linguist

The following sentence is clearly recognizable.

And this is all our extensive reading aims to do: recognize. To illuminate the parallel between Old Norse and English, clause for clause, word group for word group. To what depth? Well, somewhere before word-by-word and detailed grammatical analysis, until the Old Norse sentence structure and the correspondence appear crystal-clear. Sometimes a question nags at you, and you have a little time on your hands. So, grab a dictionary. What is the precise meaning of this suðr, place, the south of Reykjavik, or the south of Iceland, or movement, in that cardinal direction?

The class is clearly adverbial, with suðr preceded by no preposition. It can quite clearly express location, but it can also be interpreted as movement, depending on the hue given to the verb byggði: to settle (at a precise point on the map) or to go and settle (from where you dock, for example). In any case, Reykjavik is well to the south (of Iceland). We could rephrase as He settled in the south in Reykjavík or He went to settled southward in Reykjavík. This is a point of detail. The meaning and correspondence had already been achieved. Just to let you know that the depth of analysis, in truth, is up to you. Galois can't know for you when your understanding dawns. Let’s move on.

In the following, we identify our little linking words anew. In addition, we have highlighted the simple adverb of place, Þar.

We inspect the territory a little to the East, then a little to the West. The conjunction en articulates the whole. Thanks to the proper nouns and the repetition, we can clearly pinpoint our place indications.

And the two relative clauses mirror each other:

In the following sentence, perhaps the periphery reads more clearly than the center.

The time of action. Then the place, where we spot the binary structure, in the middle, ok.

This leaves

It seems that wooded renders viði vaxit. vaxit looks like a past participle. That leaves viði, and the desinence -i may appear to the connoisseur as evidence of a dative. Wooded means well covered with forest, or even closer to Old Norse, grown with forest.

Subscribe now

Among the small connecting (groups of) words (or those indicating place) that are scattered throughout what follows, let's particularly note þeir er and af því at. The first illustrates what we were saying earlier: the relative particle er is combined with the personal pronoun þeir to signify them who together. So more literally

In the second you might recognize the singular neuter dative, því, of the demonstrative pronoun, its nominative being þat. af is precisely a preposition calling for the dative, from, out of, because of. And at is the subordinating conjunction that. Literally, out of/from this, that…, which simplifies into because….

From the skeleton, everything is easily recognizable. The second independent clause.

The adverbial subordinate of cause/reason introduced by our af því at.

And the third independent clause, where the ternary structure of the object has not escaped your notice.

A last mouthful of Old Norse. Quite simple.

We recognize our previous af því, literally from this. And the subordinating conjunction at which introduces the subordinate clause object to the main clause. We have made a lot of progress.

With the translation at hand, there is nothing too difficult about extensive reading. All it takes is a little application and curiosity.

References

Þorgilsson, A. (c. 1122-1133). Íslendingabók. In F. Jónsson (Ed.). (1930). Are hinn Froðe Þorgilsson Íslendingabók. Dansk-Islandsk Forbundsfond

Ísland byggðist fyrst ór Nórvegi á dǫgum Haralds ins hárfagra, Hálfdanarsonar ins svarta, í þann tíð, at ætlun ok tǫlu þeira Teits, fóstra míns, þess manns, er ek kunna spakastan, sonar Ísleifs byskups, ok Þorkels, fǫðurbróður míns, Gellissonar, er langt munði fram, ok Þuríðar Snorradóttur goða, er bæði var margspǫk ok óljúgfróð, er Ívarr Ragnarssonr loðbrókar lét drepa Eadmund inn helga Englakonung. En þat var átta hundruð ok sjau tigum vetra eftir burð Krists, at því er ritit er í sǫgu hans.

We shall focus on this very piece of Old Norse, which although brief, is very rich in grammar. We will make three main observations, thereby highlighting three different notions.

Subscribe now

At

As we said earlier, at is the preposition which in this context means according to. But at can mean many things! Let's look closely at the entry at in the Cleasby & Vigfusson Old Norse to English dictionary. This long entry distinguishes between the cases, precisely, the grammatical cases required by the preposition, depending on use, context and meaning. Old Norse prepositions therefore call for a particular grammatical case, and sometimes several, as with at, depending on the context.

The Germanist will already be familiar with prepositions with a particular rection. Let's look at some German.

Die Feuerwehrleute gehen durch die Rauchsäule. (Firefighters make their way through the smoky dome.)
Sie kämpfen gegen das Feuer. (They're fighting the fire.)
Ein Hund läuft um die Ecke. (A dog comes around the corner.)

Durch, gegen, um always call for the accusative, and the noun phrase that follows is here correctly declined in the accusative : die Rauchsäule, das Feuer, die Ecke.

Er kommt aus dem Laden, in dem er seit mehreren Stunden eingekauft hat. (He comes out of the store where he had been shopping for several hours.)

Aus and seit both always require the dative, which the nominal groups dem Laden and mehreren Stunden bear.

Finally, two-way prepositions.

Vor dem Haus hatten sie die Geweihe eines großen Hirsches aufgehängt. (In front of the house, they had hung the antlers of a large deer.)
Die Maus rennt vor die Tür. (The mouse runs in front of the door.)

The same German preposition vor sometimes calls for the accusative, sometimes for the dative. In general, displacement demands the accusative, while the static state demands the dativ. The mouse runs, while the antlers hang.

This locative function of the dative runs through Germanic languages, and language historians date it back to Proto-Germanic. Proto-Germanic, it is hypothesized, will have completely integrated the locative case that existed in Proto-Indo-European with the dative. (Majer, 2014, p.46) A riveting chapter, and we shall soon be exploring the progression of the case system from Proto-Indo-European through Proto-Germanic to Old Norse.

Let's return to our dictionary entry. We learn that Old Norse at is followed sometimes by the accusative, sometimes by the dativ. The dative case is by far the most frequent, and Haugen (Haugen, 2015, p.140) summarizes its use as “an expression of time or place, where presence can be understood as the result of movement” (“uttrykk for tid eller stad, der nærværet kan forståast som resultat av rørsle”), which indicates a certain subset of the static state, the one arrived at after movement; interestingly, for the preposition at, the dictionary details two uses of the dative to express place, one without and the other with movement. Without movement, the preposition will be denoting presence at, near, by, at the side of, in, upon as in at ǫldri (at a banquet), at áti (at dinner), at samfǫrum ok samvistum (at public meetings/gatherings), at dómi (in court/at a legal assembly).

But the dativ is also required by the at locatif without movement, which can then translate

• towards, against, as in sendimaðrinn sneri hjǫltum sverðsins at konungi (the messenger turned the sword's hilt toward the king);

• denoting proximity, close up to as in Gunnarr kom þangat at þeim ǫrunum (Gunnar reached them there with his arrows);

• denoting a motion along, into, upon as in at ísi (on the ice) or máttu menn ganga bar yfir at skipum einum (men could walk there across by ships alone);

• hereafter somehow metaphorically, denoting an engagement as, the dictionary tells us, in ríða at hrossum, at sauðum (to go looking after/checking on horses, sheep) or

• denoting hostility as in renna at, hlaupa at (to run towards, to leap at) or

• denoting around, of clothing or the like as in vefja mǫtri at hǫfði sér (to wrap a snood around one's head)

In all cases, we sense a prevalent nuance of proximity, be it heading towards or being near.

The temporal use in dativ is rather close to the English one. Moment when something happens, point in time, sometimes more precisely, point of beginning as in at vetri (at the beginning of the winter, on the day when winter sets in).

Finally, at in dativ accompanies a number of metaphorical uses, in particular, denoting the source of a thing when learning something as in Ari nam ok marga fræði at Þuríði (Ari learned also much knowledge from Thurid) or denoting conformity - and this is the usage case in our original text - as in at ráði allra vitrustu manna (according to the counsel of all the wisest men).

The accusative usage is much less frequent, so it seems, rather for poetic and literary purposes, or in conjunction with other prepositions.

In short, Old Norse prepositions call for a specific grammatical case. Some are two-way, and may call for multiple grammatical cases, depending on meaning and usage. Let's look anew at our saga excerpt.

at ætlun ok tǫlu

ætlun and tǫlu are two feminine substantives in the dative singular, whose nominative is ætlun (reckoning, estimation) and tala (count, number) respectively.

according to the reckoning and count

margspǫk ok óljúgfróð

Let's now turn our attention to the adjectives (very) wise and truthful.

Þuríðar Snorradóttur goða, er bæði var margspǫk ok óljúgfróð

Both describe Thurid, Snorri the Goði's daughter. er is the relative pronoun expanding on Þuríðar and bæði is determiner in singular neutral, meaning both, here used adverbially. Thurid, who is both margspǫk and óljúgfróð.

Let's take a closer look at our two adjectives. The former is formed from marg and spǫk, where the former serves emphases, a bit like many- or super-. We will therefore focus on the declension of spǫk. The prefix ó- in óljúgfróður is privative while ljúgfróður means untruthful, inacurrate and the double negation gives truthful or truthfully knowledgeable. Agreed in gender and number with what or whom they refer to, Thurid, spǫk and óljúgfróð are in the feminine singular. Their form is then enough to determine that both are in the nominative, and follow the strong declension: spǫk and fróð appear nowhere else in the feminine singular declension table.

This is also indeed the case, and the declension pattern required by their function. Both are predicative adjectives, that is, adjectives appearing after a copula, here the verb to be in the preterite tense, var, and qualify the subject, functionally the relative pronoun er, which stands for Thurid. And the predicative adjective in Old Norse is, precisely, in the nominative strong declension.

For there is a weak and a strong declension of the adjective. Let's look at the weak declension of the feminine singular of spakr,

N. spaka
A. spǫku
D. spǫku
G. spǫku

and now the strong one

N. spǫk
A. spaka
D. spakri
G. spakrar

What to notice? The three oblique (i.e., non-nominative) cases of weak declension are identical. So it's the context, the number and person of verbs, the number, person, case, strong or weak pattern of substantives that helps resolve the case in situation.

All adjectives can be declined almost like spakr. Almost, because several phonetic rules apply quite regularly to certain cases, so that it is customary (Haugen) to distinguish two or three declension patterns: spakr, the most common, gamall, for adjectives ending in -al/-il/-ul or -ig/-ug and heiðinn. Of course, we shall return in detail to these important phonetic rules and to the complete adjective declension tables.

For now, let's remember that the predicative adjective follows the strong declension, that the adjective agrees in gender, number and person with the noun to which it refers, and that, so, there are two declensions of the adjective, weak and strong.

Share Old Norse Linguist

þeira

Finally, let's analyze this þeira which is somewhat resisting us.

at ætlun ok tǫlu þeira Teits, fóstra míns, þess manns, er ek kunna spakastan, sonar Ísleifs byskups, ok Þorkels, fǫðurbróður míns, Gellissonar, er langt munði fram, ok Þuríðar Snorradóttur goða, er bæði var margspǫk ok óljúgfróð

With a little experience, you can suspect without difficulty a plural genitive. The -ra/-na/-a endings are quite indicative of a plural genitive, notably in the strong declension of the adjective, and the strong and weak declensions of the noun.

at ætlun ok tǫlu, that is, according to the reckoning and count likes to call a genitive case: according to the reckoning and count of whom, a genitive construction. Let's examine our short text more closely. We start by elaguering relative clauses, which don't interest us here: this is easy, they start with the relative pronoun, er. We also eliminate þess manns, in apposition to the preceding substantive, and which the relative clause that follows completes.

at ætlun ok tǫlu þeira Teits, fóstra míns, þess manns, er ek kunna spakastan, sonar Ísleifs byskups, ok Þorkels, fǫðurbróður míns, Gellissonar, er langt munði fram, ok Þuríðar Snorradóttur goða, er bæði var margspǫk ok óljúgfróð

We now look for evidence of the genitive. Here again, a little habit informs us: -s and -ar are two recurrent genitive marks in the strong declension of the substantive. Where the neuter always takes -s, and the feminine -ar (sometimes -r plus vowel variation), the masculine takes -s or -ar on a case-by-case basis. The text is strewn with such marks, the genitives are everywhere.

at ætlun ok tǫlu þeira Teits, fóstra míns, […] sonar Ísleifs byskups, ok Þorkels, fǫðurbróður míns, Gellissonar, […] ok Þuríðar Snorradóttur goða […]

míns the pronoun possessive is also, of course, in the genitive. Then the substantive fóstra follows the weak feminine declension, with the singular genitive suffix -a. Snorradóttur and fǫðurbróður are the genitive of somewhat special nouns, belonging to a small group of kinship terms. Their three oblique singular cases make -ur. This group includes faðir (father), móðir (mother), bróðir (brother), systir (sister), dóttir (daughter). goða is the adjective in the genitive, weak masculine declension. So there you have it at last, the cartography of the genitives.

at ætlun ok tǫlu þeira Teits, fóstra míns, […] sonar Ísleifs byskups, ok Þorkels, fǫðurbróður míns, Gellissonar, […] ok Þuríðar Snorradóttur goða […]

We distinguish three noun phrases, coordinated twice by ok (and):

• Teits, fóstra míns, […] sonar Ísleifs byskups where fóstra míns (my foster-father) is a first apposition to noun Teits, sonar Ísleifs byskups (son of the bishop Ísleif) a second. In Old Norse, appositive elements agree in case with the noun they characterize or further identify. The appositional noun phrase - fóstra míns or sonar - is in genitive case because it is in apposition to something - Teits - that is in genitive case. Then, Ísleifs byskups is a further genitive construction, completing sonar.

• Þorkels, fǫðurbróður míns, Gellissonar (my uncle Thorkel, son of Gelli) is again a noun proper to the genitive, and its two appositions (Thorkel is first my uncle, second, Gellisson, that is, the son of Gelli)

• Þuríðar Snorradóttur goða (Thurid daughter of Snorri the Chieftain) where Snorradóttur goða is in apposition to and characterises Thurid

Three genitive groups, which tell us whose reckoning and count it is. What does this þeira add, then, in the genitive plural? And first of all, what is its nature?

Subscribe now

þeir(r)a is the genitive plural of the demonstrative pronoun in all genders, meaning thereby of them/their. For the record, Middle English borrowed it from Old Norse, and it derived into English their. So, litteraly, according to the reckoning and count of them, (namely) Teits, […] and Thorkel, […] and Thurid. The technical notion is that of partitive genitive or partitive apposition. This of them introduces and announces a future collection, in this case of the three proper nouns in genitive. This construction is a stylistic feature of Old Norse prose, where the demonstrative pronoun þeira precedes a collection of elements - proper nouns or something else - to create some kind of referential effect in the narrative.

We talked briefly about prepositions and their particular rection, with or without movement; the genitive case and a particular construct, the partitive apposition; and the strong and weak declension of the adjective.

In the near future, we shall be looking more closely at the intriguing ǫ (o with ogonek or caudata) for instance in sǫgu and fǫðurbróður and the related phonetic phenomenon.

Sources

Majer, M. (2014). How many grammatical cases were there in Proto-Germanic? Interpreting the Old English evidence. In E. Willim (Ed.), Young linguists in dialogue (pp. 43-50). Cambridge University Press.

Haugen, O. E. (2015). Norrøn grammatikk i hovuddrag. Novus.

Grammar may seem austere to those who have yet to discover its almost mathematical charms. We shall gladly plead in the near future in favor of a certain grammatical mastery, paving the way for the comprehension and enjoyment of texts in their original version - but we may be preaching to the converted.

Besides the grammar tables and their rigor, we shall sometimes delight in a little historical and theoretical depth, and in what follows, the subject that interests us is verbs, the Old Norse distinction between weak and strong verbs.

If you are a native English speaker, or if English is second nature to you, you may forget in everyday life that you are constantly manipulating verbs that are sometimes weak, sometimes strong. And yet, you never make mistakes.

Yet let's take a closer look.

Subscribe now

The verb whose infinitive is to ride conjugates as follows, where we show on a line the indicative present and past tenses followed by the past participle (which forms the compound tenses).

ride - rode - ridden

Let's further observe, for example, to take, to freeze and to grow.

take - took - taken
freeze - froze - frozen
grow - grew - grown

These verbs are strong, and contrast with weak ones such as to cook or to play.

cook - cooked - cooked
play - played - played

The respective characteristics become clear. The weak verb seems to form the indicative past tense and the past participle by simply adding the suffix -ed. The present indicative is the infinitive in all persons but the third singular, which we know appends a terminal -s or -es.

Well, this is almost the full picture. The indicative past tense or past participle sometimes adds only a -d or a -t to the infinitive, as in to love or to dream respectively.

love - loved - loved
dream - dreamed or dreamt - dreamed or dreamt

as per a British English grammar.

Our strong verbs, for their part, vary the vowel, that is their distinguishing feature.

ride - rode - ridden
take - took - taken
freeze - froze - frozen
grow - grew - grown

More precisely, the present indicative retains the infinitive form and vowel. The past indicative alters the vowel, and in took, omits the final -e from the infinitive. The past participle also alters the vowel, adds an -n and in ridden doubles the consonant. These vowel variations seem to be quite erratic, but we remember noticing certain trends in elementary school, an o - e - o pattern in

grow - grew - grown
throw - threw - thrown

or ea - o - o in

break - broke - broken
speak - spoke - spoken

These are the remnants of a rather distant phonological process, already underway in Proto-Indo-European. Our reconstruction of Proto-Germanic, the most recent ancestor common to all Germanic branches, indicates a fairly regular system of strong verbs, comprising roughly seven coherent classes, each characterized by a vowel alteration pattern. These seven classes are well preserved in the first batch of Germanic languages: Gothic, Old English, Old Norse, Old High German, and fade thereafter. (Ringe, 2017, p.262-263)

Let's observe the remnants of the same phenomenon in modern German, for example

finden - fand - fanden - gefunden (to find)
zwingen - zwang - zwangen - gezwungen (to coerce)
glimmen - glomm - glommen - geglommen (to glow)
beginnen - begann - begannen - begonnen (to begin)
schwimmen - schwamm - schwammen - geschwommen (to swim)
trinken - trank - tranken - getrunken (to drink)

where both the 3rd. singular and 3rd. plural forms are given for the preterite (past) indicative. Those specimens can be associated with the Proto-Germanic 3rd class, just like the following modern English verbs

shrink - shrank - shrunk
begin - began - begun
swim - swam - swum
drink - drank - drunk
cling - clung - clung
bind - bound - bound
win - won - won

and all show a similar vowel shift i - a - u (ou), with substitution of -o- in the past participle for geschwommen or begonnen, in the past indicative (preterite) and past participle for glomm - glommen - geglommen or win - won - won, with substitution of -u- (or -ou-) in the past indicative for clung and bound.

With these examples, we have pretty much mapped out the modern German and English 3rd class of strong verbs, or more precisely, a subset of it. A very lucid mind will have noticed a common trait in all these verbs that make i - a - u or approximately.

finden - zwingen - glimmen - beginnen - schwimmen - trinken

all share a certain sameness: their stem is nasal, the stem vowel is followed by an -n, exceptionally by an -m. Incidentally, our perverted -o- cases are precisely such -m exceptions. The same goes for English

shrink - begin - swim - drink - cling - bind - win

There is even a little more in common: the nasal (-n) is (almost) always followed by a consonant.

finden - zwingen - glimmen - beginnen - schwimmen - trinken
shrink - begin - swim - drink - cling - bind - win

Incidentally, we now see that our German -o- exceptions match exactly the infinitive stems that repeat the nasal consonant: glimmen - beginnen - schwimmen.

Happy with our findings, let's ascend back to the Proto-Germanic ancestor, following the example of this 3rd class, first (nasal) subset.

The pattern of this subclass was present *iN, past indicative singular *aN, default past (e.g., plural) *uN, past participle *uN while the root always ended in a consonant. (Where the nasals writes with a capital letter.) Some twenty verbs belonging to it could be reconstructed (Ringe, 2017, p.269), including

*finþaną, *fanþ, *fundun, *fundanaz (to find)
*drinkaną, *drank, *drunkun, *drunkanaz (to drink)
*brinnaną, *brann, *brunnun, *brunnanaz (to burn, intransive)

whose descendants we are delighted to trace:

finþan - fanþ - funþun (Gothic)
finna - fann - fundu - fundinn (Old Norse)
findan - fand - fundon - funden (Old English)
findan - fand - funtun - funtan (Old High German)

and of course the modern

finden - fand - fanden - gefunden (High German)
find - found - found (English)
vinden - vond - gevonden (Dutch)
finne - fant - funnet (Norwegian)
etc.

The same goes for the following examples, with the Gothic

drigkan - dragk - drugkun - drugkans
brinnan - brann

the Old Norse

drekka - drakk - drukku - drukkinn
brenna (~ brinna) - brann - brunnu - brunninn

the Old English

drincan - dranc - druncon - druncen
birnan - barn - burnon - burnen

and the Old High German

trinkan - trank - trunkun - gitrunkan
brinnan - brann - brunnun - gibrunnan

Here at last is the Proto-Germanic ancestor of our exceptional verb, featuring a nasal geminate

*swimmaną - *swamm - *summun - *summanaz

with numerous descendants, including the Old Norse

svimma - (swim) - svamm - summu - symmi - summinn

Incidentally, we are here using Haugen's notation for strong verbs

infinitive - (indicative present 3rd pers. sg.) - indicative preterite 3rd pers. sg. - indicative preterite 3rd pers. pl. - (conjunctive preterite 3rd pers. sg.) - supine

where the infinitive is also the indicative present 3rd pers. sg. and represent the present stem. (Haugen, 2015, p.102) Supine denotes, as far as Old Norse and Icelandic are concerned, the nominativ and accusative neuter of the past participle.

This has been a great exercise in genealogy, which we shall pursue further in the near future. In short, the Proto-Germanic strong verb system inherits vowel variations stemming from a phonological process of vowel gradation, called ablaut, that was active in Proto-Indo-European. Those vowel variations outlines seven Proto-Germanic classes of strong verbs, initially rather regular, later fading away with the emergence of the descendant languages, but still traceable in modern Germanic languages. Characteristically, strong verbs have several distinct stems, namely for the indicative present tense, the indicative past tense singular and plural, and the past participle. The set of these radicals presents a vowel variation, the typical pattern of which links the verb to one of the strong verb classes.

Share Old Norse Linguist

What about weak verbs? Let's look briefly at the German for to cook or to love.

lieben - liebt - liebte - liebten - geliebt (to love)
kochen - kocht - kochte - kochten - gekocht (to cook)

Those are German weak verbs, and as with English weak verbs, the stem vowel is not altered by conjugation. The ending -te/-ten is added to the present tense stem to form the preterite, -t to form the past participle. As in English, German weak verbs form their preterite and past participle by means of a dental suffix, that is to say, some /t/ or /d/ sound or similar: this, anew, dates back to Proto-Germanic times.

Proto-Germanic originally inherited the strong Proto-Indo-European verbs, but the ablaut or vowel gradation process that formed them is no longer productive in Proto-Germanic. (The vowel variation “rule” is no longer active, so if a "Proto-German" coins a new verb, that verb is not subject to the rule.) As a result, almost all new verbs in Germanic languages are weak verbs, and most of the original strong verbs decay into weak verbs by analogy.

What characterizes all weak verbs, Ringe tells us, is the common formation of the past participle, of a default finite past tense (all tenses, modes and persons), with the exception of a distinct singular indicative past tense. More specifically, the past participle of weak verbs always has a stem vowel *-a- or *-ō-, and is formed with the suffix *-da- (with a few exceptions). The past tense suffix begins with the dental obstruent *-d-, with the suffix and endings following a well-defined pattern. (Ringe, 2017, p.280) (To level out the terminology, a conjugated verb is formed by a stem, a suffix (here, the past tense suffix begins with a *-d-) and an ending typical of the mode, tense and person.)

Ringe distinguishes four major classes of weak verbs in Proto-Germanic, plus three unclassifiable verbs, relics of Proto-Indo-European and said to have a simple thematic present tense. (Ringe, 2017, p.262-263) We shall gladly explore further the typologies and mechanisms of Proto-Germanic weak (and strong) verbs at a later date, but for now, and to strictly serve our purpose, we are content to observe a largely represented weak class, class II as per Ringe.

Verbs in this second class have a suffix *-ō- in the present tense, *-ōd- in the past tense, and *-ōda- to form the past participle. Take *kalzōną, for example. The third person makes *kalzōþi in the present indicative, *kalzōdē in the past indicative, while the past participle is reconstructed as *kalzōdaz. It gives Proto-West Germanic *kalʀōn, Old English ceallian and English to call. On the other hand, Old Norse kalla.

Verbs of this weak class, therefore new verbs, which Proto-Germanic created, often derive from a noun or adjective. *laþōną (to invite) from *laþō (invitation) and *fiskōną (to catch fish) from *fiskaz (fish) make very regularly

*laþōþi - *laþōdē - *laþōdaz
*fiskōþi - *fiskōdē - *fiskōdaz

After this necessary archaeological preliminaries, we turn to Old Norse. Only three weak verb classes remain, and the typology is very clear. We follow Haugen. (Haugen, 2015, p.103)

The first Old Norse class is the heir of the second Old Proto-Germanic weak verb class we just visited. Kalla belongs therefore to it, the same goes for kasta, so Haugen says, kasta-class. It is characterized by a past tense suffix composed of a linking vowel, -a-, followed by the dentalsuffix.

We have, using the form

infinitive - (present indicative) - preterite indicative - (preterite subjunctive) - past participle

kasta - (kastar) - kastaði - (kastaði) - kastat
kalla - (kallar) - kallaði - (kallaði) - kallat

The remaining two Old Norse classes outlined by Haugen do not seem to us to correspond bijectively to the Proto-Germanic classes delineated by grammarians.

Verbs from the second Old Norse class of weak verbs as per Haugen, the telja-klasse, are characterized by a preterite suffix with no linking vowel, formed by the dentalsuffix alone, and a vowel shift from the preterite to the present stem, e.g.,

gremja - (gremr) - gramdi - (gremdi) - gramt
telja - (telr) - taldi - (teldi) - talt
flytja - (flytr) - flutti - (flytti) - flutt)

Some inherit from the Proto-Germanic class I, globally endowed with suffixes having *-i- or *-j- in the present and preterite tenses, e.g. *gramjaną (to get angry) - *gramiþi - *gramidē - *gramidaz giving the Old Norse gremja.

However, from *hauzijaną, which clearly belongs to the same class I as *gramjaną - *hauzīþi - *hauzidē - *hauzidaz derives the Old Norse heyra, a clear representative of the third class under Haugen. This third class, the so-called dǿma-class, is characterized by a past tense suffix with no linking vowel, formed by the dentalsuffix alone. In contrast to the second class, there is no stem vowel shift, e.g.,

heyra - (heyrir) - heyrði - (heyrði) - heyrt
dǿma - (dǿmir) - dǿmdi - (dǿmdi) - dǿmt
fǿra - (fǿrir) - fǿrði - (fǿrði) - fǿrt

What about Proto-Germanic class III and IV verbs? Some class III verbs, such as *þulāną and *wakāną, give partial representatives of the døma-klass

þola - (þolir) - þoldi - (þøldi) - þolat
vaka - (vakir) - vakti - (vekti) - vakat

as a vowel change affects the preterite subjunctive like in the second class, whereas the past participle is formed with a linking vowel as in the first class. And as for the Proto-Germanic class IV, says Ringe, it fuses with classes II and III while evolving into Old Norse.

We shall soon continue to excavate the mysteries of these typologies.

A sharp mind, however, will have twitched. We said that strong verbs are characterized by a pattern of vowel variation, and yet our second class of weak verbs has one too, like

telja - (telr) - taldi - (teldi) - talt

Good point. The key is that this is not the same type of vowel variation. The vowel shift of the weak verb telja is not a Proto-Indo-Enropeen remnant (of the ablaut process), but the result of a process active in Old Norse. Because telja injects a -j-, its stem -a- shifts to -e-. This is a phonological i-umlaut, and we shall have ample opportunity to discuss it afresh.

Old Norse thus has strong and weak verbs. Strong verbs are characterized by a set of stems with vowel variations, while weak verbs form the past tense and the past participle with a dentalsuffix, preceded or not by a linking vowel. We are now well equipped to detect the verbs when viewing a new sample from the sagas, as we shall soon be doing anew in the Grammar by sagas and Norse practice series.

References

Ringe, D. (2017). From Proto-Indo-European to Proto-Germanic (2nd ed.). Oxford University Press.

Haugen, O. E. (2015). Norrøn grammatikk i hovuddrag. Novus.

Wiktionary, Appendix:Proto-Germanic_verbs https://en.wiktionary.org/wiki/Appendix:Proto-Germanic_verbs

Subscribe now

Let’s observe closely. This is from the first chapter of Íslendingabók.

En þat var átta hundruð ok sjau tigum vetra eftir burð Krists, at því er ritit er í sǫgu hans.

The following is the Islandic equivalent.

En það var átta hundruð og sjötíu vetur eftir burð Krists, samkvæmt því sem ritað er í sögu hans.

The closeness is uncanny. Let us look at the Norwegian equivalent.

Og det var åtte hundre og sytti vintre etter Kristi fødsel, ifølge det som er skrevet i hans saga.

Then the Swedish one, and finally the Danish one.

Och det var åtta hundra och sjuttio vintrar efter Kristi födelse, enligt vad som står skrivet i hans saga.

Og det var otte hundrede og halvfjerds vintre efter Kristi fødsel, som det er skrevet i hans saga.

We are not very far away either, but maybe a little further.

As you can appreciate, there is a difficulty inherent in the “closeness” in question. We can spontaneously detect and gauge the proximity of languages. In the present case, to say with a simple glance of neophyte, that Old Norse is closer to Icelandic than it is to the three continental ones, without being far from the latter. If, however, we try to rationalize this impression, to explain in detail the "components" or "modalities" of this “closeness”, we may be left speechless. How close, precisely, are both languages, and close how? It takes a bit of thinking.

Galois dreams of alchemizing a precise, very mathematical notion of distance between languages. A machine which, given the universal corpora of any two languages, engineers their objective score of similarities, in all respects, according to a secret and exact formula. Until then, we happily spend our time researching among humans, intuiting and marveling at the closeness of Old Norse to his close heirs.

Our dream of a distance is revealing of a number of problems in this comparison undertaking. Let's patiently consider anew our sentence from Íslendingabók.

En þat var átta hundruð ok sjau tigum vetra eftir burð Krists, at því er ritit er í sǫgu hans.

By the way, the following is an English translation:

And it was eight hundred and seventy winters after Christ's birth, according to what is written in his saga.

We are at the very beginning of the first chapter of Íslendingabók. We are learning about the temporal circumstances of Iceland's first colonization, as well as the famous and reliable sources relied on by the author, Are Torgilsson Frode. Our Norwegian translation was the following:

Og det var åtte hundre og sytti vintre etter Kristi fødsel, ifølge det som er skrevet i hans saga.

Yet, alternative propositions are very valid. For instance, this works equally well.

Og det var åtte hundre og sytti vintre etter Kristi fødsel, slik det står skrevet i hans saga.

Let’s compare.

Let's take a short detour into grammar. at is an Old Norse preposition that can mean many things. In the present case, it is followed by the dative, and is probably well translated by according to, entry C. VI. in the Cleasby & Vigfusson Dictionary. því is indeed the dative of the neutral singular demonstrative pronoun, which is declined as follows:

Nom. þat
Acc. þat
Dat. því
Gen. þess

er is the relativ pronoun that the Norwegian translates here as som. at því er is therefore literally: according to that which, corresponding well to the Norwegian ifølge det som. Thus, the first Norwegian translation of the relative proposition in Old Norse, after the comma, is virtually the word for word translation of the Old Norse.

The second renders at því er by slik det. This does not alter the meaning, but melts the according to that which into a slik det which literally rather means such as that. So, *according to that which is written in his saga becomes in the second Norwegian translation something approaching such as it stands (står) in his saga.

We can see that the first translation, ifølge det som, is obviously closer to the Old Norse than the second, slik det.

Here is the problem with our distance we seek to illustrate here. Translation is not an exact science. For the same Old Norse sentence, we can produce one Norwegian translation that corresponds word for word, and another a little further away. Distance is not a function of the language pair alone, but fluctuates with translation. It can be objected that these variations are minor, and do not alter the conclusion immediately intuited, that Old Norse is much closer to Icelandic than to the other three. The fact remains, however, that the possibility of an exact calculation of a closeness is very much challenged by this wide variability in translation. Our case here is simple: translation, in general, is interpretation, and creation. The interplay is therefore important.

Another obvious challenge in defining a distance is the variability of the languages themselves. Great authors can be said to have a language of their own. As we shall see in an upcoming Norse chronicle, linguistic tendencies, the style of the writers is a crucial clue when attempting to trace the authorship of the sagas. Of course, the official language tends to freeze written usage, but even so, it can still fluctuate to some extent geographically or sociologically. A slang or dialect word is woven into the essay. And of course, language evolves over time. The reference dictionary sells a new edition every Christmas. Languages let themselves be anglicized or overall influenced, they sharpen, they broaden, they jargonize, and so on. Some languages are evolving rapidly (globalization), while others seem to be freezing millennia-old characteristics (insularity).

Icelandic, then, is one of those languages that preserves, as if in ice, the remnants of old Norse.

Subscribe now

A modern Icelandic reader can very well understand our Old Norse sentence:

En þat var átta hundruð ok sjau tigum vetra eftir burð Krists, at því er ritit er í sǫgu hans.

when the following is our Islandic equivalent:

En það var átta hundruð og sjötíu vetur eftir burð Krists, samkvæmt því sem ritað er í sögu hans.

Let's focus on the first proposition.

En þat var átta hundruð ok sjau tigum vetra eftir burð Krists

En það var átta hundruð og sjötíu vetur eftir burð Krists

The syntax ( the structure of the sentence) is rigorously identical, as is the word order, every Old Norse word finding an Icelandic equivalent whose similarity is sometimes complete. A minor exception is the number sjau tigum in two words (seventy) which becomes sjötíu in Icelandic. eftir burð Krists remains perfectly unchanged. Interestingly, this points to a very analogous declension paradigm. Krists in burð Krists is indeed the genitive of Old Norse Kristr, and of Icelandic Kristur.

Kristr presents the -r mark of the nominative of certain masculine substantives. The interspersed u in Kristur most likely appeared by epenthesis, a phological process present in a large number of languages across multiple language families, in particular the Germanic languages, and which calls for some clarification.

Generally speaking, epenthesis is the exact opposite of elision, a word perhaps better adopted by everyday language. Elision removes a sound, epenthesis adds one, vowel or consonant, either spoken or written.

In Old Norse (and Old Icelandic), the final -r was probably syllabic in certain circumstances, for example when it followed a single consonant (Þráinsson, 2017). What is a syllabic consonant? Typically, syllables require a vowel as their nucleus. However, certain consonants can take on vowel-like qualities and serve as the syllabic core. In the International Phonetic Alphabet, this is a overstroke (or understroke) diacritic - a small dot or line above (below) the sound in question. This concerns for instance m, n and l in some pronounciation of some English words such as

little in its UK [ˈlɪtʰɫ̩] or US [ˈlɪɾɫ̩] or [ˈɫɪɾɫ̩] pronounciation
prism [ˈpʰɹ̠̊ɪzm̩]
listen [ˈlɪs.n̩]

In short, we can assume that the suffix -r marking the Old Norse nominative Kristr was such a vocalic consonant, a consonant that is a syllable of its own. This is also the case with the substantive dalr (valley) or the adjective latr (lazy), generally, according to certain rules, notably when this final -r follows a simple consonant. (Þráinsson, 2017)

Since Modern Icelandic does not allow syllabic consonants, a /u/ - the slashes denote the sound - was inserted between the /r/ and the preceding consonant. Here is our epenthesis, which has a systematic allure in the evolution from Old to Modern Icelandic

dalr → dalur
latr → latur
Kristr → Kristur

A sharp mind will note that in Kristr, this -r follows a group of two consonants, -st. The rule, we learn from Þráinsson (2017), admits a few exceptions: the epenthesis is not realized when the stem ends in a high vowel followed by /l,n/ or in /r, s/ or in certain clusters of consonants such as /ss, gl, gn/. Well, we still don't know much more about the Kristr case. This u-epenthesis began at the end of the thirteenth century. We can imagine a very similar, somewhat generalized phenomenon occurring at some point. Not as generalized as in Farroese, however, who also realized this u-epenthesis in the three cases of exception just mentioned. We can illustrate the discrepancy, from Old Icelandic to Modern Icelandic and Modern Faroese respectively, still with Þráinsson (2017), when the stem is ending in a high vowel followed by /l,n/

stóll (Old Icelandic) → stóll (Modern Icelandic)
stóll → stólur (Modern Faroese)

fínn → fínn (Modern Icelandic)
fínn → fínur (Modern Faroese)

or when the stem is ending with /r, s/

íss → ís (Modern Icelandic)
íss → ísur (Modern Faroese)

stórr → stór (Modern Icelandic)
stórr → stórur (Modern Faroese)

But what happened to our -r mark of Old Norse nominative singular stóll, íss, fínn? Simply, some kind of progressive assimilation - the preceding vowel “contaminates” the following vowel. Note that the same assimilation occurs in Icelandic in the case “stem is ending in a high vowel followed by /l,n/” making

*stól-r → stóll
*fín-r → fínn

In the cases “stem is ending with /r, s/” or “stem certain clusters of consonants such as /ss, gl, gn/” another process occurs: the -r - or its assimilation into the preceding consonant - is simply deleted.

*ís-r → ís
*stór-r → stór

We shall come back to assimilations in Old Norse at length, in the Grammar dive rubric.

Let's look anew at our very close sentences, in Old Norse and Icelandic.

En þat var átta hundruð ok sjau tigum vetra eftir burð Krists, at því er ritit er í sǫgu hans.

En það var átta hundruð og sjötíu vetur eftir burð Krists, samkvæmt því sem ritað er í sögu hans.

þat notably becomes það, the symptom of a more general shift in from Old to Modern Icelandic, where a voiceless dental stop - /t/ in þat - weakens to the voiced dental fricative in unstressed positions - /ð/ in það. It occurs as part of a larger pattern where “the voiceless stops are drawn toward spirantization”, especially in final word position, which are more vulnerable to weakening. Spirantization, which an example will clarify sufficiently for now, generally takes place in two stages, with stops first undergoing affrication, such as [t] → [ts], before affricates become pure fricatives, [ts] → [s]. (Goblirsch, 2001)

This phenomenon explains a great many things. While English water maintains the proto-germanic /t/, the latter spirantalizes into /s/ in German, making Wasser. The process is only halfway through in German Zunge where the inital /ts/ affricate (written Z in German) stems from the voiceless dental /t/ in English tongue. This High German consonant shift only affected the southern parts of the West German dialect continuum, around the 6th-7th centuries CE, leaving English unaffected and still bearing the marks of the distant past. The trend seeks to materialize across the entire Germanic spectrum, Goblirsch teaches us. The shift observed in Icelandic stems from an identical progression, developed independently, at around the same time, in the North Germanic branch. Which leaves plenty of work for the near future: spirantisations, consonant shifts, and the general phonology of Old and Modern Icelandic. We also leave to the trial of your patience the much awaited comparison of Old Norse and Old Icelandic.

The Old Norse relative pronoun er became the Icelandic sem, which this time resembles the Norwegian som. í sögu hans remains very close to the original í sǫgu hans.

Finally, if we compare our Old Norse sentence

En þat var átta hundruð ok sjau tigum vetra eftir burð Krists, at því er ritit er í sǫgu hans.

with the Norwegian equivalent,

Og det var åtte hundre og sytti vintre etter Kristi fødsel, ifølge det som er skrevet i hans saga.

we notice a preservation of structure and form that is nonetheless far-reaching.

With this we shall, in closing, emphasize the obvious benefits of mastering Norwegian, or a closely related modern language, for a well-mastered exegesis of Old Norse.

Sources

Þráinsson, H. (2017). U-umlaut in Icelandic and Faroese: Survival and death. In C. Bowern, L. Horn, & R. Zanuttini (Eds.), On looking into words (and beyond): Structures, relations, analyses (pp. 99-112). Language Science Press.

Goblirsch, K. G. (2001). The Icelandic consonant shift in its Germanic context. Arkiv för nordisk filologi, 116, 117-133.

Subscribe now

As this is our first “lesson” in extensive reading, we shall give special attention to the method. Typically, pedagogues of linguistics, eager for seriousness and to adequately typologize everything, define extensive reading as the exercise that seeks a context-based general understanding of the text under study, more than complete comprehension and detailed analysis: in short, a focus on quantity and fluency. We have our own reading of extensive reading, and this exercise illustrates our approach. It cannot be ruled out that, over time, we further refine and improve on how we proceed.

We will read and examine the beginning of the first chapter, entitled Frá Íslands byggð, which can be translated as About the settlement of Iceland.

This is the very first paragraph, our first "long" text, and its English translation. Reading the translation is allowed at this stage, we shall see what our Icelandic saga is actually about.

Ísland byggðist fyrst ór Nórvegi á dǫgum Haralds ins hárfagra, Hálfdanarsonar ins svarta, í þann tíð, at ætlun ok tǫlu þeira Teits, fóstra míns, þess manns, er ek kunna spakastan, sonar Ísleifs byskups, ok Þorkels, fǫðurbróður míns, Gellissonar, er langt munði fram, ok Þuríðar Snorradóttur goða, er bæði var margspǫk ok óljúgfróð, er Ívarr Ragnarssonr loðbrókar lét drepa Eadmund inn helga Englakonung. En þat var átta hundruð ok sjau tigum vetra eftir burð Krists, at því er ritit er í sǫgu hans.

Iceland was first settled from Norway in the days of Harald Fairhair, son of Halfdan the Black. According to the reckoning and count of Teit, my foster-father (whom I knew to be the wisest man), son of Bishop Isleif, and my uncle Thorkel Gellisson, who remembered far back, and Thurid daughter of Snorri the Chieftain, who was both very wise and reliably knowledgeable, this was when Ivar son of Ragnar Lothbrok had Saint Edmund, King of England, killed. That was eight hundred and seventy winters after the birth of Christ, as it is written in his saga.

We learn more about the historical background to the first settlement of Iceland by Norwegians. The author evokes his sources, who are also his famous ancestors.

The Old Norse has only two sentences, the first is very long, and this is our first working material. It ends with some considerations about a certain Eadmund inn helga Englakonung. So, we can see that the English translation corresponding to this first long sentence stops on the considerations on Saint Edmund, King of England, who was killed.

So we start at the beginning and try to map the first Old Norse proposition, highlighted in blue, with its English equivalent.

How do we know when we have a proposition? A little experience tells us. More precisely, we quickly identify byggðist as a verb (see Grammar by sagas here). We glance at the surrounding words, in the highlighted section, and find no other finite verbs. The immediate continuation, í þann tíð, introduces some indication of time, while at seems to us to be a preposition, inserting an adjunct of some kind. A typical Old Norse proposition will have a subject, a verb and a object. As we explained earlier, despite the possible debate among linguists, the structure subject, verb, object, in this order, has a high probability. But it is not the only possible one. The sophisticated inflectional system dispenses with strict word order. There are also a number of situations in which the subject, verb or object is elided; impersonal constructions; and short sentences in which the verb is in infinitive form. The grammarian Haugen calls these "situations" reformulations: the original "normal" proposition - subject, verb, object - has been reformulated respectively into a proposition whose subject, verb or object is elided, or into an impersonal proposition, or into a short infinitive phrase. (Haugen, 2015)

The association with English is made particularly simple by the abundance of proper nouns, e.g. Nórvegi for Norway, the similarity in word order, and of course, similarities in form, e.g. fyrst for first, dǫgum for days.

The rest of the identification process is far less straightforward. Let’s turn back to our probable indication of time.

The noun tíð is no mystery, a little imagination speaks for at that time. Now, the very first time the notion of time is expressed in the English equivalent, is with when:

A whole story of sources and ancestors is interspersed, whose correspondence with the English can be guessed thanks to the anchors provided by proper names.

We also note the similarity of the relative clauses: the Old Norse one is introduced by er. Knowing that ok is the coordinating conjunction and helps, of course.

The incise clause can therefore be delimited,

which leaves the piece that interests us now, expressing this when.

Here too, er introduces a relative proposition. In this time, when or as it rendered here, this was when. So, when what? When Ivar son of Ragnar Lothbrok had Saint Edmund, King of England, killed.

An attentive mind may object that the time adjunct formed by what has just been underlined in blue belongs to the first proposition, the main one, and that the relative proposition when Ivar son of Ragnar Lothbrok had Saint Edmund, King of England, killed is enshrined in this main one. This is a very valid opinion. In this reading, the main Old Norse proposition covers at least what is highlighted in blue in the following

and as we shall see in a moment that the central incise is a vast adjunct (of manner), it actually extends to the entire piece quoted. For our progressive breakdown, which seeks to understand Old Norse-English matches, we do need to draw subsets, as we have done. A first, main position, which could function independently. Its time adjunct, which contains a relative proposition, and now, the big incise starting with the preposition at.

We have already spotted the final relative proposition, in which the author panders to his ancestors, both margspǫk and óljúgfróð. We can spot another one easily, as it begins here too with the subordinating conjunction er.

And a third, for the same reason.

The rest of the Old Norse incise is clearly comparable to the English, full of proper nouns, and without technical difficulties.

We take a breath of fresh air, and resume with the beginning of our very long proposition (or, if you prefer, with our independent proposition, the one that has been stripped of its furthest adjuncts).

Everything revolves around the verb: the subject Ísland, the place adjunct (expressing provenance) ór Nórvegi, the time adjunct fyrst, the time adjunct á dǫgum Haralds ins hárfagra. Here, even more precisely, we have the correspondance

and then that of the genitives, the noun complements

The genitives are themselves completed by an apposition, Hálfdanarsonar ins svarta, of Harald Fairhair, which is also in the genitive.

Now you have the gist.

With these Russian doll-like associations of word groups, we have tried to retrace what might be going on in the mind of an extensive reader. It is less a question of reading "without fully understanding", that is to say, skimming, than of methodically finding the translation anew - a historical translation right in front of your eyes, the one obtained with a modern generative A.I. tool, or the one you have in mind because you are familiar with the passage.

Ísland byggðist fyrst ór Nórvegi á dǫgum Haralds ins hárfagra, Hálfdanarsonar ins svarta, í þann tíð, at ætlun ok tǫlu þeira Teits, fóstra míns, þess manns, er ek kunna spakastan, sonar Ísleifs byskups, ok Þorkels, fǫðurbróður míns, Gellissonar, er langt munði fram, ok Þuríðar Snorradóttur goða, er bæði var margspǫk ok óljúgfróð, er Ívarr Ragnarssonr loðbrókar lét drepa Eadmund inn helga Englakonung.

Reread the old Norse patiently, and note, probably, that you "understand". What understanding means, is probably: knowing how to decompose quite precisely, and each piece, matching it to something you already know.

And to conclude the opening paragraph of the first chapter of Íslendingabók, read, likewise, the second sentence:

En þat var átta hundruð ok sjau tigum vetra eftir burð Krists, at því er ritit er í sǫgu hans.

parallel to its translation:

That was eight hundred and seventy winters after the birth of Christ, as it is written in his saga.

sǫgu is indeed saga. The circumstances of this vowel change, from a to the ǫ-vowel, also called o with ogoneko, or o caudata, will be the subject of much comment later.

Here, for saga and far north enthusiasts, the second paragraph of the first chapter of Íslendingabók, to be read in closer details later.

A Norse man named Ingolf is truthfully said to have gone first from there to Iceland, when Harald Fairhair was sixteen winters old, and for a second time a few winters later. He settled in the south at Reykjavik. Ingolf's Headland is the name given east of Minthak's Eyr, where he first came ashore, and Ingolf's Mountain west of Olfossa River is where he later claimed his property

References

Haugen, O. E. (2015). Norrøn grammatikk i hovuddrag. Novus.

We have no regrets about these essays; they document our practical philosophy on the matter exemplarily. From their earlier edition, we revised them slightly to emphasize their genuine purpose: not an exhaustive course in Old Norse the school way, but the substantiation, by way of Old Norse, of our general approach to learning languages like a machine, whose foundations we present here, and of which Learn Finnish like a machine is the most recent illustration.

We attempted to typologize the numerous Old Norse Linguist essays into different rubrics. But, ultimately, it’s all about practicing language learning the machine way. We will be happy to further expand the series over time, although we must admit that we are nowadays busy with our other Far North affairs.

Share Old Norse Linguist

Galois is convinced that languages, both ancient and modern, are learned through tireless and repeated exposure to textual data and their translation. Not just by having them run before your eyes, but by working with them, understanding them, and replaying them over and over again. Here, learning an ancient language is akin to learning a modern one: you forget, you relearn, until you’re exhausted.

In the Read Old Norse like a machine rubric, we try to convey how we practice that core exercise of ours.

Machines learn in exactly the same way: through exposure to massive textual data, bilingual data from which they are trained to identify correspondences, word block by word block, at different scales.

But we are no machine. We don't reach the realms of the very, very large numbers that allow for straightforward inference of all things grammar, lexicon or syntax. Therefore we would do well to integrate, at a certain point in the learning process, the one-time reading of a grammar book. And to endeavor to pay attention, in the course of our extensive readings, to the grammar at work.

Essays in the Grammar dive rubric dispense Old Norse grammar in bite-sized chunks, based on the great texts. Here, we map out and make accessible to the international reader some fundamental bits of Old Norse grammar as they are explained in excellent Old Norse grammar textbooks available in Scandinavian languages.

In a Multilingual affairs rubric, we indulge in our favorite exercise: the comparison, the game of Spot the difference, between Old Norse and other closely related languages, ancient and modern. It’s about translations studies and comparative linguistics, from Old Norse to modern Scandinavian (North Germanic) languages: Norwegian, Swedish, Danish, Feroen, Icelandic; but also modern West Germanic languages: German, Dutch, English; and some of their ancestors: Old Icelandic, Old English. Here we discuss language history and geography, language sociology, idioms and dialects. We talk about archaisms and modernisms, and bridge the gap between the old and the new.

Some essays ponder broader Linguistics topics. We marvel for example at how the future tense makes its way into grammar in different languages.

Finally, further essays talk about miscellaneous topics. Centuries of philological work have sought to shed light on the great sagas' manuscripts: eminently, John Megaard traces the authorship of the works and the circumstances of their creation. Or, the future of the work of mind in the age of intelligent machines.

Be regular, and above all, make good use of your time.

Subscribe now

Subscribe now

We've said that the dentalsuffix -ð- helps mark the preterite. Let's try to find all the finite verbs in this short passage, i.e. all verbs conjugated in the present or preterite tense. By skimming through, we can spot a number of forms that might be suspected of having a dentalsuffix.

Ísland byggðist fyrst ór Nórvegi á dǫgum Haralds ins hárfagra, Hálfdanarsonar ins svarta, í þann tíð, at ætlun ok tǫlu þeira Teits, fóstra míns, þess manns, er ek kunna spakastan, sonar Ísleifs byskups, ok Þorkels, fǫðurbróður míns, Gellissonar, er langt munði fram, ok Þuríðar Snorradóttur goða, er bæði var margspǫk ok óljúgfróð, er Ívarr Ragnarssonr loðbrókar lét drepa Eadmund inn helga Englakonung. En þat var átta hundruð ok sjau tigum vetra eftir burð Krists, at því er ritit er í sǫgu hans.

Ingólfr hét maðr nórrænn, er sannliga er sagt, at færi fyrst þaðan til Íslands, þá er Haraldr inn hárfagri var sextán vetra gamall, en í annat sinn fám vetrum síðar. Hann byggði suðr í Reykjarvík. Þar er Ingólfshǫfði kallaðr fyr austan Minþakseyri, sem hann kom fyrst á land, en þar Ingólfsfell fyr vestan Ǫlfossá, er hann lagði sína eigu á síðan.

byggðist, munði, byggði, lagði are indeed preterite verbs.

bæði i “er bæði var margspǫk ok óljúgfróð” is something else. How do you know? Well, first of all, in this proposition, er is clearly the finite verb. This is the present tense of the verb to be, which you could learn right now.

ek em
þú ert
hann/hon/þat er
vér erum
þér eruð
þeir/þær/þau eru

What is more, bæði is a frequent form you might want to be able to recognize. It is far from being a verb. bæði is the neuter plural form nominative and accusative of the determiner báðir, meaning both.

Back to our verbs with an obvious dentalsuffix -ð-. byggði and byggðist differ only in that the second carries the suffix typical of mediopassive, an Old Norse form expressing, among other things (here) the passive.

munði, byggði, lagði are the preterite of weak verbs, while some are strong. -i is the common 3rd person preterite inflection for these weak verbs, which form the preterite with dentalsuffix -ð-, -d- or -t-, depending on the final stem sound.

lagði forms from the infinitive leggja (identical to the 3rd person plural present), which demonstrates a characteristic vowel shift from the root vowel -a- (in lag-) to -e-. This shift occurs due to i-umlaut, a systematic sound change triggered by the presence of the -j- elementone of the fundamental phonological processes in Old Norse, affecting numerous verbs and other word classes and one of the fundamental phonological processes in Old Norse, affecting numerous verbs and other word classes.

And as far as munði is concerned, it is of a rather special kind. Its preterite is that of weak verbs, here the dentalsuffix is clear. But its present tense is shaped like the preterite of strong verbs!

In this instance, muna does in the present tense:

ek man
þú mant/manst
hann/hon/þat man
vér munum
þér munuð
þeir/þær/þau munu

What features of the conjugation of strong preterite verbs stand out here? First of all, a vowel change in the singular with respect to the infinitive - not triggered by the presence of a -j- (or anything else) in the continuation of the word in Old Norse, but the residue, the heritage of an earlier vowel change - potentially triggered according to other rules by certain combinations of consonants and vowels, but the trace of which has been lost as letters have “fallen out” over time. Next, the typical endings of strong preterite verbs, namely -∅, -t, -∅, -um, -uð, -u.

We shall pause here for a moment, and resume a little later our reading of the tales of Iceland, which, we shall learn, was first installed in Norway during the reign of Harald Fair-hair, son of Halfdan the Black.

1. Frá Íslands byggð.

Ísland byggðist fyrst ór Nórvegi á dǫgum Haralds ins hárfagra, Hálfdanarsonar ins svarta, í þann tíð, at ætlun ok tǫlu þeira Teits, fóstra míns, þess manns, er ek kunna spakastan, sonar Ísleifs byskups, ok Þorkels, fǫðurbróður míns, Gellissonar, er langt munði fram, ok Þuríðar Snorradóttur goða, er bæði var margspǫk ok óljúgfróð, er Ívarr Ragnarssonr loðbrókar lét drepa Eadmund inn helga Englakonung. […]

Let's focus on the very beginning.

Ísland byggðist fyrst ór Nórvegi á dǫgum Haralds ins hárfagra, Hálfdanarsonar ins svarta, í þann tíð, […]

The finite Old Norse sentence contains a verb in either present or preterite tense. Old Norse allowed considerable variation in word order because its rich inflectional system marked the grammatical function of words through case endings already. Rigidly assigning a place in the sentence to each function was then redundant with the form, and the language admits a certain flexibility.

Old English had a similar inflectional system, but as English lost its inflections during the Middle English period, and word order gained in importance, to show grammatical function. English had to develop a more rigid word order pattern to maintain clear meaning.

However, this flexibility of Old Norse is not infinite. There are two opposing views in the controversy over the configurational or non-configurational nature of Old Norse. The traditional Norwegian view considers it non-configurational with Subject-Verb-Object (SVO) as the most frequent surface order. Let's clear up the jargon. Non-configurational implies a free word order, where constituents, that is words or phrases that belong together syntactically, are (potentially) discontinuous, that is, separated by other elements in the sentence. An example in English: Switch the light on, where the verbal form switch on is spread around the object the light. A non-configurational language relies therefore on primarily on its case system to mark the function, and make sense. In short, the Norwegian view sees Old Norse as having a free word order with nevertheless a typical SVO pattern.

The traditional Islandic view considers Old Nores configurational: there's an underlying basic word order (SVO), and if there are variations, they follow specific syntactic rules: well-determined movement operations explain the apparent freedom.

Back to our Íslendingabók.

Once we're a little more experienced, we'll be able to spot the finite verb in an instant, i.e. the verb conjugated in the present or preterite (past) tense.

Ísland byggðist fyrst ór Nórvegi á dǫgum Haralds ins hárfagra, Hálfdanarsonar ins svarta, í þann tíð, […]

byggðist has a rather special -st form. -ð- is very indicative of a preterite form, as dental suffix is part of the Germanic innovation for forming past tenses. Weak verbs in Old Norse (there are weak and strong verbs in Old Norse) form their preterite with such a dental suffix, which realizes as -ð-, -t- or -d- selon les cas.

-st ending marks the mediopassive voice, which developed from a fusion with the reflexive pronoun accusative, sik. As the name suggests, mediopassiv has something to do with passiv. It can actually express reflexive actions ("I dressed myself"), reciprocal actions ("We hugged each other"), passive meaning ("It has been done") or impersonal actions (“It is raining”). Here, byggðist means "was settled" or "became settled", with clear passive meaning. Our verb breaks down into

• bygg- (stem)

• -ð- (dental suffix, weak verb preterite marker)

• -st (mediopassive ending)

So that's the grammatical analysis from the very beginning.

Ísland byggðist fyrst ór Norvegi
Iceland was (first) settled from Norway

Ísland: subst. nom.
byggðist: vb. pret. 3. sg. mediop.
fyrst: adv.
ór: prep.
Norvegi: subst. dat.

The nominative (nom.) is the case of the subject, while the dative (dat.) of Norvegi is governed by the preceding preposition.

We've learned a lot of new things and will be back at it soon.