It is unquestioned that all languages change over time, but do languages change for some purpose? If someone proposes that languages do in fact shift in some particular direction, towards some particular goal, they are proposing a teleological explanation. A teleological explanation is one that takes a purpose or goal to be inherent. For example, hypothetical principles of language change such as “Morphosyntactic change occurs in order to make processing easier”, or “Sound change occurs in order to make articulation easier” are teleological in nature.
Generative linguists have not traditionally placed substantial importance on language change. Typical generative perspectives on what is interesting about language constrains inquiry mostly or entirely to the synchronic domain. A commonly held belief within generative linguistics is that the structure of one’s grammar is set in place during first language acquisition, and therefore changes in grammar can only happen in transmission, during early childhood. Since the Chomskyan paradigm limits the window for language change to a short period in early life and assumes a universal set of domain-specific principles that strongly constrain linguistic structure, the diachronic domain becomes rather unimportant. From this perspective, any language change that happens will be constrained by the universals, so the universals should be the focus of inquiry rather than the changes themselves. The universals are non-teleological as formulated, in that they are presented as synchronic constraints, not as desired or optimal states.
Mostly outside of the generative tradition, other researchers view the diachronic domain as crucial in explaining synchronic grammars and view adult grammars as malleable. Adult grammars must be able to change over time, otherwise one must ignore empirical evidence for the gradual nature of language change, and the demonstrable relationship between diachronic processes and synchronic grammars (some important works relevant to the relationship between synchrony and diachrony include Blevins 2004, Mielke 2008 and Bybee 2010). Diachronically oriented theories may or may not be teleological, but I would argue that a teleological position is undesirable because it does not seem to be the case that all languages shift towards some particular goal or set of goals (see Lass 1997: 340-352). Rather, we find a wide range of phenomena in grammars, some that might fit a proposed goal and others that would run counter to it. Teleological theory in diachronic linguistics has a similar burden to certain proposed properties of Universal Grammar, in that exceptions are bound to exist. Take, for example, the pervasive idea of universal phonological features. Miekle (2008: 3) demonstrates that none of the three most popular universal feature theories, Preliminaries to Speech Analysis (Jakobson, Fant & Halle 1951), The Sound Pattern of English (Chomsky & Halle 1968), or Unified Feature Theory (Clements & Hume 1995), were able to characterize more than 71% of the observed sound classes in his sample of 628 languages, and 24% were uncharacterizeable in any of the three theories. My point is to highlight the fact that even though empirical investigation reveals pervasive patterns in language, demonstrating true universality is difficult or perhaps impossible. If one is uncomfortable setting aside counterexamples, an alternate approach is needed. This, in my view, should be both diachronic and non-teleological in nature. Such a theory would view structure as emergent, rather than falling out from synchronic universals or diachronic teleologies, and would have the potential to account for both common and rare linguistic structures.
What is meant here by “emergent” structure is linguistic structure that comes about as a result of facts about speech perception, articulation and cognition, and how they affect language change. Linguistic emergence is comparable in important ways to evolution in the biological sciences (though of course they cannot be equated). Where an organism finds itself in an ecological environment, a grammar finds itself in what might be called a physio-cognitive environment, by which I intend to refer to the cooperation of cognitive processing (including speech perception) and vocal tract anatomy and physiology. Where species tend to change in particular ways depending their natural surroundings, grammars tend to change in particular ways depending on a range of cognitive and articulatory factors involving vocal tract aerodynamics, speech perception, and domain-general cognitive processes such as memory, frequency effects (chunking), and analogy. In both cases, change is constrained by environmental factors, which results in a synchronic structure. Of course, there are strong similarities in the “environment” of any grammar, since all typically-developing humans share the same cognitive capacities and have similar vocal tract structure and physiology, and as a result we see some strong general tendencies between grammars cross-linguistically. However, just as with biological species, there is potential for many varied types of grammars to emerge in a given environment, though the basic mechanisms driving change will stay the same. This argument, which I will refer to as the “emergentist” position, is non-teleological in nature, as is evolutionary theory when properly conceived.
As it were, some in the biological sciences lament the use of teleological statements (see, for example, Hanke 2004), such as “the beak of this bird species changed so that they would have an easier time eating grubs,” precisely because that does not capture the nature of emergence. Rather, it would be better to say something like “the members of this bird species that developed this beak mutation propagated, because it happened to allow them to eat more grubs”. The difference may seem subtle, but it is the difference between setting up a goal, such that the birds were somehow destined to eat more grubs, and setting up an instance of emergence, such that those birds that happened to have this mutation propagated. The mutation itself was not directly motivated, but was merely one possibility given the genetic code of that species and its environment at the relevant point in history. For something analogous in language take the cross-linguistically common process of word-final or phrase-final devoicing. Some factors at work in final devoicing processes involve the pre-pausal nature of final positions in speech. Right before you pause is a time at which it is natural for the vocal folds to begin relaxing in preparation for the coming pause. The trans-glottal airflow is then insufficient to maintain the same level of voicing with the abducting vocal folds, and can lead to total devoicing (Blevins 2004: 104). As language is transmitted among members of a speech community, the tendency to devoice pre-pausal segments can become phonologized, such that final devoicing becomes part of the grammar. This comes about because of an interaction between vocal tract physiology (naturally reduced vocal fold vibration before a pause), and cognition (the detection and phonological incorporation of the natural devoicing tendency). Of course final devoicing is not guaranteed to happen (not all languages do this), but it is common cross-linguistically – far more common than a phonological process of final obstruent voicing, for example, which is virtually unattested (Blevins 2004: 192). In this example, facts about the aerodynamics of speech in the context of language use help to explain the cross-linguistic pattern, while still acknowledging the languages that do not follow the trend. This is non-teleological in that the change doesn’t happen as a result of some intrinsic goal, but rather occurs as a greater-than-chance event given the nature of pre-pausal positions in speech, the way that our vocal tracts function, and the cognitive processes by which phonological patterns are established in grammars.
As a final note, I’d like to address a criticism one sometimes hears regarding usage-based or emergentist frameworks in linguistics. The criticism is that they are untenable because they allow for circular reasoning, or in other words, that they are tautologies. The complaint is that when a theory posits that observed phenomena are due to some factors shaping language over time, one uses the very phenomena observed as evidence that these factors exist. An analogous criticism is levied against Darwinian evolutionary theory. The argument is such: if survival of the fittest drives evolution, then the one who survives is the fittest. However, the way one knows who is the fittest is by observing who survived, and therefore there is no basis for determining fitness independent of survival (De Cuypere 2008: 183-184). This argument is not very convincing because fitness can be observed empirically. It seems that polar bears are adapted to the arctic, and cacti to the desert, for example. The situation for language is a bit more complicated, since we are not dealing with a range of discrete natural environments to observe, but rather general aspects of speech and cognition. Careful observation, greater knowledge of linguistic typology, and experimentation to confirm physio-cognitive influences on language change are needed to further research on the emergence of language structure.
Blevins, Juliette. 2004. Evolutionary phonology: the emergence of sound patterns. Cambridge; New York: Cambridge University Press.
Bybee, Joan L. 2010. Language, usage and cognition. Cambridge; New York: Cambridge University Press.
Chomsky, Noam & Morris Halle. 1968. The sound pattern of English. New York: Harper & Row.
Clements, G.N. & Elizabeth Hume. 1995. The internal organization of speech sounds. In John Goldsmith (ed.), The handbook of phonological theory, 245–306. Cambridge, MA: Blackwell.
De Cuypere, Ludovic. 2008. Limiting the iconic: from the metatheoretical foundations to the creative possibilities of iconicity in language. Amsterdam; Philadelphia: John Benjamins Publishing Company.
Hanke, David. 2004. Teleology: the explanation that bedevils biology. In John Cornwell (ed.), Explanations: Styles of Explanation in Science, 143–155. New York: Oxford University Press.
Jakobson, Roman, Gunnar Fant & Morris Halle. 1951. Preliminaries to speech analysis: the distinctive features and their correlates. Cambridge, MA: MIT Press.
Lass, Roger. 1997. Historical Linguistics and Language Change. New York: Cambridge University Press.
Mielke, Jeff. 2008. The Emergence of Distinctive Features. Oxford; New York: Oxford University Press.
Comments by Kevin Hughes