Recently, I’ve been hunting for a good overview of Algonquian finals and their semantics. It’s easy to find high-level descriptions, like this one which describes the high-level structure of Mi’gmaq verbs, but I’ve not yet found a comprehensive analysis. But based on what I’ve read, the similarity to other light verb constructions (e.g. in Basque, Urdu, …) and Australian classificatory verbs seems striking. The only difference is that very productive compounding and incorporation in Algonquian has combined the light verb and its complement into a single verbal word while retaining the productivity of the system and also making it an obligatory construction.

I’m sure this is far from a unique insight, I just haven’t noticed anyone labelling finals as light verbs before.

My Understanding of Algonquian Initials, Medials, and Finals

My understanding of the Algonquian verb is as follows, from here. (Almost?) all verbs contain both an initial and a final, which form a bipartite stem which can be separated by medials. The finals contribute:

• Transitivity / argument structure marking
• Classification of verbal argument(s) as animate or inanimate
• Contribution of concrete meanings from a fairly limited class, often classificatory or based on manner or instrument of action

The finals also crucially seem to be the head, in terms of determining the grammatical function of the word. An initial combined a verb final is a verb, and an initial combined with a noun final is a noun.

The verbs finals seem to divide into a set of abstract finals which mostly mark argument structure and animacy, and concrete finals which also add more detailed event classification or manner information. Some examples from the same website:

Full word	Initial	Final	Translation
a’tugw-e-t	a’tugw	e	S/he tells a story
a’tugw-ew-atl	a’tugw	ew	s/he tells a story to him/her

The medials which can occur between these two elements are basically either incorporated nouns or classifiers of the verb’s arguments.

A Simple Example of Obligatory Light Verb Constructions: Basque

Starting with a very simple system of obligatory light verbs, we have Basque, which has now reached the point where almost all verbs require an auxiliary. A small number expressing common meanings still retain a synthetic conjugation, including come, go, wander, bring, have, be, take, and know, but the rest require support from a finite verb to form a predicate.

As the auxiliary has become an obligatory part of the clause, the choice of auxiliaries has also contracted. In modern Basque, the choice is basically one of transitivity and agentivity: transitive verbs and some agentive intransitives take one set of forms based on the verb ‘have’, and the rest take a different set based on the verb ‘be’. The tense and aspect are expressed by a combination of suffixes on the non-finite verb, and the tense and mood of the auxiliary.

Compare the following examples of an intransitive and a transitive clause, and note the difference in auxiliary choice, and the TAM suffixes on the non-finite verb:

A complicating factor is that forms identical to transitive auxiliaries can also be selected to have a slot for gender agreement with familiar second persons. The following is an illocutive form:

The Basque system is similar to the system of Algonquian finals in the sense that the auxiliaries form an obligatory part of the clause, but the auxiliaries have become so semantically bleached in the process that they are an impoverished system by comparison.

Australian Verb Classification: CVs and IVs

Many aboriginal Australian languages also have (almost) obligatory light verbs, but with more complex semantics. The choice of light verb / auxiliary reflects not just transitivity, but also the general event type. The typical terminology is to call the light verbs independent verbs (IVs), and their complements coverbs (CVs).

There is a clear cline in Australia in the phonological and semantic integration of CVs+IVs, from languages with very large sets of IVs, which often occur alone, to languages with a smaller IV class and CV+IV as the typical clause type, and from languages where the CV and IV are separate words to ones where they lose their independence and merge.

Jaminjung is a good example of the case where the class of IVs is relatively small and the CV+IV construction is highly frequent. According to Eva Schultze-Berndt’s grammar, the following verbs have finite forms:

• -yu BE
• -muwa HAVE
• -irdba FALL
• -arra PUT
• –ijga GO
• -ruma COME
• -uga TAKE
• –anJama BRING
• -unga LEAVE
• -arrga APPROACH
• -wardagarra FOLLOW
• -mili/-angu GET/HANDLE
• -ma HIT
• -ina(ngga) CHOP
• -inama KICK/STEP
• –ijja / -yaluga POKE
• -wa BITE
• -wardgiya THROW
• -irna BURN
• -irriga COOK
• –yu(nggu) SAY/DO
• -ngarna GIVE
• -yungga TAKE AWAY
• –ngawu SEE
• -minda EAT
• -(ma)linyma MAKE

There are also a few low frequency or marginal verbs:

• -garra EXCRETE
• -yangma FEAR
• -malangawu HEAR
• -warrwa SWEAR AT
• -yima TELL A LIE
• -inijba DO BY MAGIC
• -ngardgani BE SICK
• -manka BE ANGRY
• -yangi BE

The IV+CV combinations are not completely predictable, but when used in combination with an IV, the general criteria for selection of an independent verb include:

• Aktionsart (stative, atelic dynamic, telic)
• Causation, transitivity and control
• Stationary vs locomotive action
• Instrument and manner, including:
  • Prolonged (atelic) contact vs telic contact event
  • Nature of impact or force (hit, chop, kick, poke, bite, act with body weight, by heat, …)

Compare the following examples from Schultze-Berndt’s grammar. The first is of an IV alone, whereas the second is of an IV+CV combination:

Schultze-Berndt states that the CV and IV form a tightly-knit unit prosodically, with primary stress on the CV and secondary stress on the IV.

Crucially, in CV+IV constructions, the choice of IV is not fossilised or fixed for each CV, but varies with the event being described. Compare the following examples, which all contain the CV “jab”:

Australian Polysynthesis

The interesting thing is that some Australian languages go further along this road. The Daly River languages seem undoubtedly polysynthetic, and the structure of the verb is vaguely reminiscent of the Algonquian verb. All examples here are from The Oxford Handbook of Polysynthesis.

Firstly, the IV and CV have merged into a single word in that order, as in the following example from Murrinhpatha:

“Shove” here would correspond to the final in Algonquian and “tell” to the initial. The order is reversed, but the general division of meaning seems vaguely similar. Secondly, many of these languages have noun incorporation, and in some of them, the incorporand occurs between the two parts of the predicate:

This order of lexical verb, incorporand, and light verb is the exact inverse of the Algonquian verb structure.

Conclusion

I end where I began: it seems to me that the bipartite structure of the Algonquian verb is very reminiscent of light verb constructions elsewhere, which may become obligatory or almost obligatory. What distinguishes Algonquian, and the Australian languages with this structure, from languages elsewhere is the richness of the light verb class.

In most Eurasian languages which have gone down this road, grammaticalisation has also involved the class of light verbs shrinking to the point where they mark little except transitivity / argument structure. We see this in Basque, where the auxiliary verbs are almost semantically empty. This has also happened in Australian languages, although there are languages like Jaminjung which have heavily grammaticalised the construction and still retain a closed class of 20-50 IVs.

I also find it very interesting that the semantics of the concrete finals seem to have something in common with the semantics of Australian IVs. My impression that both are composed of a small set of universally very high frequency verb meanings (e.g. be, have, go, …) AND a set of verbs primarily describing or classifying manner or instrument of action or the force dynamics of the event. The specific result of the action is primarily described by the CV / initial.

I do not know how many productive finals Algonquian languages typically have, but my impression is that the number may be significantly larger than Australian languages at a similar stage of grammaticalisation. There are Australian languages with a closed class 100-200 IVs, but these are typically languages where IVs alone are still frequent in speech and the CV+IV construction has lower frequency / is less obligatory. Algonquian seems to have arrived at a situation which the CV+IV equivalent is basically obligatory, while maintaining a class of productive finals much larger than the 20-30 of Jaminjung. And this situation seems to have been stable since Proto-Algonquian.

Therefore, the real exceptionality of the Algonquian bipartite stem seems to be primarily in the size of the class of classificatory verbs / independent verbs / light verbs.