Cherokee/On Polysynthesis

What's the deal with "polysynthesis"? edit

As you may know, Cherokee is described as a polysynthetic language, which to the layperson means something like “verb inflection will be the bane of my existence.” I entreat you to approach this subject with all the curiosity you can lure out, and nuance.

First, what is synthesis?

Synthesis occurs in many languages and is essentially combining morphemes into a single word. Morphemes are units of meaning. In English, for example, one could analyze the word “wanted” as having two morphemes, “want” and “ed.” The first denotes the general meaning of the verb, i.e “to desire” or “to wish to have or obtain,” etc. while the second denotes that the action or state of wanting occurred in the past. English doesn’t employ synthesis all that much in the grand scheme of things, so it is not considered synthetic enough to get the label “synthetic” slapped on the whole language.


In contrast, you have a so-called polysynthetic language like Cherokee. Polysynthesis refers to the fact that statistically speaking, words in Cherokee are composed of a higher number of morphemes than do those languages not considered polysynthetic. Fortunately, both in general and in the case with Cherokee, not all types of words are composed of multiple smooshed-together morphemes. Underived nouns in Cherokee, which make up a substantial chunk of Cherokee nouns, for instance, comprise only single morphemes and only undergo synthesis in the formation of locative constructions. Possession for underived nouns, for example, is simply accomplished the same way as in English, by using the noun in conjunction with a separate word indicating the possessor.

Verbs, however, are the domain where polysynthesis is pervasive in Cherokee. In a typical Cherokee verb inflection, the first morpheme could indicate the subject (the doer of the verb), the second would be the verb root itself, and then you would have a morpheme indicating how the event of the verb fits into the timeline (e.g. “the action is taking place now,” or “the action just happened a moment ago”) and finally a suffix further specifying one of several subtypes related to time.

So ... what exactly is the problem with polysynthesis?

Why does this system trouble people so much? After all, if you’re combining morphemes together, why can’t you just memorize the individual morphemes as if they were standalone words (even though you know technically they can’t stand alone), like how you can do with an analytic—meaning “non-synthetic”—language? Then you could just say those little “words” in a specific order, and voilà!—you have a fully inflected Cherokee verb! And then, if this is indeed possible, perhaps you could even reason with yourself that the only reason these morphemes are considered word parts and not actual words may be that historically people just never wrote the word parts with spaces between them. If they had written these morphemes with spaces in between, then perhaps nowadays we would not be using “polysynthetic” to describe Cherokee. In other words, could the label of polysynthesis be a guise, essentially a just-so descriptor based on arbitrary historical conventions of thinking about and writing the language?

There are two answers to this, first a practical one, and then a nerdier approach that associates the term “polysynthesis” with a rather specific phenomenon that happens to be in Cherokee, thus justifying its usage.

First, the practical answer.

While it is possible to have a language that is called polysynthetic arbitrarily, there are a couple of good reasons to consider Cherokee morphemes within verbs just word parts. First, these morphemes always occur in a particular order. Words are typically understood to have at least some flexibility in positioning. However, this is not the case in Cherokee verb morphology. Rather, morphemes in the Cherokee verb strictly stay in their one allowable position relative to other types of morphemes in the verb. For example, the reflexive morpheme always follows the pronoun morphemes if it is used, and the verb root always comes after the reflexive morpheme. This inflexibility means that the “words” are essentially “glued” to specific slots in the verbal template. Since they cannot freely move in this template system, there is no strong incentive to analyze the morphemes as independent words.

Secondly, almost certainly as a result of inflexible ordering, most parts of the Cherokee verb have multiple forms differing slightly based on which sounds (phonemes) are found around them, or based on the usage of the verb in the larger context of the clause. One can see that these sound changes entail “words” affecting the pronunciation of neighboring words, and hence it may be easier to make sense of the language’s structure by encapsulating all parts that affect each other as a single unit. This encapsulation allows a learner, speaker, or linguist to maintain the distinction between two levels of grammar—the nitty-gritty, lower-level, sound-change-abounding realm of morphology, involving interactions between morphemes within the context of the word; and the higher-level arena of syntax, which plays with choice, and notably, order, between words. By calling the Cherokee verb a “word,” the morphological details of its inflection do not intrude into syntax, which helps keep the complexities of each system of grammar contained and thus contextualized within its own set of rules.

Second, the "nerdier" answer.

The second, “nerdier” answer to the usefulness of the polysynthetic label is that in linguist Mark Baker’s 1996 book The Polysynthesis Parameter, he makes the case that so-called polysynthetic languages, identified at the surface level by such phenomena as “sentence words,” where seemingly whole clauses can be condensed into a single word, often have behind their deeper structure a property that would be useful to identify as the true origin of polysynthetic-like surface structure. This property is whether in general, a dependent unit of meaning is required to be marked (i.e. identified) on the structure that it depends on. It will turn out that Cherokee grammar mostly aligns with this property being present.

In a more concrete way as it pertains to the verb, the object of the verb is “dependent” on the verb, as is the subject. For context, in linguistics, it is generally taken that in any clause, the verb is the central element. The verb can have dependent arguments that further specify the meaning the speaker intends. For example, in the sentence “A boy sees a cat,” the central element is the verb “sees,” and this verb allows for two arguments, its subject “a boy” and its object “a cat.” In this English example, however, notice that only a reference to the subject is made on the verb through the use of the -s suffix to indicate a third person singular subject. English syntax then makes it clear that the subject is specifically the boy.

Why isn’t the cat marked in some way on the verb? It seems that in English, it is simply not required for the verb to make some reference to its object. This is not the case in Cherokee, however. In the Cherokee sentence “achûja àgowhtíha weésa” (one formulation of “a boy sees a cat”), the à- prefix on the verb contains the meaning of a third person singular subject (achûja, “boy”) and a third person singular object (weésa, “cat”).

This is an illustration, in plainer terms, of how requiring the marking of dependents results specifically in one characteristic of Baker’s conception of polysynthesis, polypersonal agreement. Stated most simply, in a polysynthetic language the verb marks for subject and object.

The other characteristic of Baker’s polysynthesis is productive noun incorporation, which allows objects to actually fuse with the verb root. This might allow for a verb root that meant “cat-seeing.” In related languages like Mohawk, this process is indeed productive, but in Cherokee it is no longer productive, with only a small proportion of verbs displaying evidence of historical noun incorporation. As such, Cherokee doesn’t allow for “sentence words” containing the object beyond a few fossilized objects such as garments. However, because the language exhibits polypersonal agreement and historical noun incorporation, the polysynthetic label still makes sense.

The Pros of Polysynthesis!

With this context, I suggest that the reason so many learners fear the word “polysynthetic” is that they have not properly contextualized polysynthesis in the larger scheme of grammar, or that they see the details of the system, particularly the sound changes, too readily as difficult, complex, and unnecessary rather than phonetically optimized, rich, and orderly. It’s both and neither, actually, but as a learner, which side would you rather dwell on?

I’ll leave you one positive feature of polysynthesis as food for thought: the strict order of morphemes leaves little room for ambiguity in the construction of verbal inflections. The formula is unchanging and the slots are already set up for you, each in its allocated position waiting to be filled. If you follow the formula to the tee, there’s no room to err! Now, your job is to learn all the morpheme options for the individual slots and how they change their sound to blend in smoothly among their neighboring morphemes. Once you become proficient in this process, you will already be able to express so much of the semantic load of your thoughts using Cherokee! Do not fear polysynthesis—embrace it instead!