Language complexity as an evolving variable
Sampson, G., Gill, D. & Trudgill, P. (Eds.) 2009. Language complexity as an evolving variable. Oxford: Oxford University Press.
Reviewed by Christina Behme.
Language evolution has received considerable interest in recent decades (e.g., Hauser, 1996; Dunbar, 1997; Deacon, 1997; Knight et al., 2000; Givon & Malle, 2002; Wray, 2002; Christiansen & Kirby, 2003; Wildgen, 2004; Burling, 2005; Tallerman, 2005; Johannson, 2005; Cangelosi, 2006; Hurford, 2007; Lyon et al., 2007; Tomasello, 2008; Botha & Knight, 2009). In the process many widely held assumptions about language have been re-evaluated. This volume is a the result of a 2007 workshop, debating and challenging the longstanding linguistic axiom that all languages and language varieties are and always have been similar in complexity. The contributions (19 chapters) span a wide variety of disciplines presented by experts from around the globe. In the first chapter co-editor Geoffrey Sampson explains why for most linguists the assumption that language complexity does not vary among different human languages or over time has been taken for granted even though there was never "offered much justification for it" (p. 1). The most influential schools in 20th century linguistics (descriptivists and generativists) differed considerably in many of their fundamental assumptions about the nature of language but essentially agreed on the complexity axiom. The reasons for this agreement vary from the belief that language is a genetically determined part of human biology to the assumption that "all languages have about equally complex jobs to do" (Hockett, 1958, p. 180) and 'political' concerns about the perceived correlation between language complexity and the 'sophistication' of the speaker. Yet, empirical evidence gathered from numerous human languages seems to point towards considerable differences in complexity. Such differences exist between present-day and 'ancient' human languages, between speakers of a given language, over the lifetime of individual speakers, between 'old' and 'new' languages and between languages spoken by large and small communities. Before turning to some of this evidence it is important to stress a point not explicitly made by Sampson. The (alleged) fact that language is a genetically determined part of human biology alone could not account for invariance in language complexity. After all, traits like eye-, hair-, or skin-colour are also genetically determined, yet they vary considerably across the species and even across local populations. The claim by many defenders of linguistic nativism (e.g., (Chomsky, 1966, 1980, 1986, 2002, 2005, 2007; Crain, 1991; Smith, 1999; Hauser, Chomsky & Fitch, 2002; Matthews, 2006; Lightfoot, 2006; Jackendoff, 2007; Isac & Reiss, 2008; McGilvray, 2009; Hornstein & Boeckx, 2009) is stronger. For them language is an essential part of human nature: "[t]he language faculty ... constitutes what is distinctive about human nature" (McGilvray, 2009, p. 50) and any variability in such a species defining trait would be ruled out. Therefore, challenges to the invariance of linguistic complexity axiom introduced in the present volume would not only constitute the melting of a linguistic 'iceberg that has floated into warm waters' (p. 18) but also cast doubt on the essentialist view of linguistic nativism. To be clear, alternative views (e.g., language as non-essential biological trait or as product of cultural evolution) do not necessitate that language complexity is an 'evolving variable' but they allow for this possibility.
Given the length limitations of this review I will only focus on two of the individual chapters. The choice of these is based on personal interest and does not reflect on the quality of the remaining chapters. In chapter 12 (Individual differences in processing complex grammatical structures) Ngoni Chipere provides empirical evidence challenging the "widespread assumption that all native speakers posses a uniform underlying competence to process complex grammatical structures" (p. 178). Contrary to the Chomskyan dictum that all "individuals in a speech community have developed essentially the same language" (Chomsky, 1975, p. 11), Chipere claims that researchers have shown that individuals "vary in almost every aspect of grammatical performance: in phonology, morphology, the lexicon, and syntax" (p. 179). Traditionally, nativists have attempted to account for those observable differences by reference to differences in the performance systems while insisting that the underlying 'competence' is uniform across the species (e.g., Chomsky, 1975, 1980, 1986; Crain, 1991; Smith, 1999), Chipere suggests testing this hypothesis. If differences in performance systems (e.g., working memory) cause differences in language output, then specific memory training should improve performance. Grammar training on the other hand should have no effect because it is not possible to improve the innately determined competence. However, if differences in performance are caused by differences in grammatical competence, then grammar training (and not memory training) should improve performance. Based on these assumptions Chipere designed experiments to test the impact of grammar- and memory-training on sentence recall and comprehension, in subjects with significantly different language performance. He found that memory training only increased recall while grammar training increased recall and comprehension and concluded that "native speaker variations in understanding complex sentences arise from differences in grammatical competence and not in working memory capacity" (p. 190). This finding challenges the assumption that competence is based on a genetically determined universal grammar (e.g., Chomsky, 1975, 1980, 1986, 2007; Smith, 1999; Matthews, 2006; McGilvray, 2009) which is invariant across the species and cannot be affected by 'training'.
In chapter 15 (An interview with Dan Everett) Geoffrey Sampson engages in a clarifying discussion about "[o]ne of the most startling recent events in linguistics" (p. 213), the report of a human language (Piraha) that lacks such allegedly 'essential' features as grammatical recursion (e.g., Hauser, Chomsky & Fitch, 2002), quantifier words (all, some, most, etc.) and number (one, two, etc.) words. The discussion touches on the most obvious question (whether or not Everett got the facts about Piraha right) and on the implications of Everett's findings for research methodology, anthropological fieldwork, adequate politico-cultural evaluation of speakers of 'primitive' languages, measurement of cultural effects on grammar, evaluation of cultural progress vs. cultural decay and translatability of concepts from one language into another. Everett stresses the importance of empirically testable hypotheses for linguistic theorizing and rejects "the introspective work characteristic of...'armchair linguistics'" (p. 220). Further, he contrasts the close scrutiny his work on a largely unknown, 'exotic' language has received to the general acceptance of work on well studied and 'easily translatable' languages. Specifically, the different interpretation of some facts about Piraha in 1986 and 2005 publications has been interpreted as a worrisome contradiction by critics (e.g., Nevins, Pesetsky & Rodrigues, 2009; Chomsky, 2007). Yet the potential implications of Everett's research for established theories have been downplayed (e.g., Fitch, Chomsky & Hauser, 2005; Chomsky, 2007) and comparably substantial changes to those theories have been hailed as "substantial break from earlier generative grammar" (Chomsky, 1986, p.7), "radical reformulation of the theory" (Chomsky, 1995, p. 378), "continuous series of theories" (Kasher, 1991, p. 122), "important developments" (Catell, 2006, p. 181), "shifting frameworks" (Collins, 2008, p.4), "successive simplifications in the theory" (Lasnik, 2005, p. 79), "profound rethinking" (Newmeyer, 2005, p. 57) or "striking manifestations of...intellectual vigor" (Smith, 1999, p. 86). One can remain skeptical about the correctness of Everett's interpretation of Piraha data but his point about our tendency to be "not as demanding on the factual claims that support our preexisting ideas as we are on the claims that contradict them" (p. 220) should not be dismissed lightly. Thus, the most important implication of the work on Piraha might not be that this language disproves Universal Grammar (as suggested by the popular press) or challenges the principle of invariance of language complexity, but that it can serve as a reminder that we ought to remain open to the possibility of revising 'established' theories in light of new data.
The remaining chapters of Language complexity as an evolving variable provide additional empirical evidence undermining the dogma that language complexity is a universal constant. David Gil demonstrates that the complexity of many existing languages by far exceeds the functional requirements of effective communication. Thus, there is a wide gap between existing language complexity and the complexity required by the 'jobs that languages have to do' and grammatical structure cannot be a "straightforward tool for the communication of pre-existing messages" (p. 32). Östen Dahl suggests that from the fact that "it is not possible to assess the global grammatical complexity of a language" (p. 63) it does not follow that grammatical complexity is invariant between languages. He shows that there is a significant difference in morphological complexity between Swedish and Elfdalian and that it is questionable whether this difference will be compensated by syntactic complexity. Peter Trudgill discusses the impact of social factors (e.g., degree of contact/isolation, denseness/loseness of social networks, community size) on language complexity and also stresses that currently existing languages are only a small (and possibly atypical) subset of all human languages. It becomes evident from these and other chapters, covering a diverse range of human languages, that another Chomskyan dictum (we can form "strong empirical hypotheses concerning universal grammar, even from the study of a particular language" (Chomsky, 2006, p.100)) needs to be revised. Finally, given the wide acceptance of the invariance of language complexity axiom and the equally widely held belief that the 'Chomskyan Revolution' elevated linguistics to the level of an 'exact, rigorous science' (e.g., Chomsky, 1957, 1975, 1986, 2006; Smith, 1999; McGilvray, 2009), it seems reasonable to assume that precise criteria for measuring language complexity exist and are rigorously applied. Yet the observations of several authors make it evident that reason and linguistic axioms are not always directly correlated. The problem of measuring complexity has many faces: some linguistic structures exist "whose complexity is hard to quantify" (Bisang, p. 48), "overall complexity of the syntactic component of a language is not easily measured" (Dahl, p. 62), some proposed complexity measures (e.g., implicational hierarchies) have only been investigated for "a limited number of grammatical phenomena" (Miestamo, p. 96), "[c]omplexity levels are hard to measure and weigh" (Nichols, p. 120), "it is important to devise better and more comprehensive measures of complexity" (Nichols, p. 121), "the focus on inflection [as easily accessible measure for complexity]...can distract us from the nature of grammatical complexity in a broader sense" (McWhorter, p.161), "it is impossible to define a single measure of overall complexity for a language's grammar" (Deutscher, p. 250), "there is no clear metric of overall complexity" (Hawkins, p. 253). And, as if this were not bad enough, one author observes also disagreement about what complexity is and how it should be defined and suggests that "ranking grammars according to overall complexity is problematic" (Hawkins, p. 267) until we have reached an agreement about what needs to be measured. The attempts of some contributors to develop and apply measuring criteria for specified aspects of complexity (e.g., Miestamo, pp. 80-97; Szmrecsanyi & Kortmann, pp. 64-79; Nichols, pp. 110-125; Sinnemäki, pp. 126-140; Maas, pp. 164-177; Karlsson, pp. 192-202) are commendable but much work remains to be done until these methodological problems are resolved.
Overall Language complexity as an evolving variable is an interesting, well researched and engagingly written contribution. It presupposes some familiarity with the subject matter but only very few of the chapters are written at a technical level that is accessible only to experts. The editors acknowledge that while virtually all contributors "take variation in language complexity for granted" (p. 269) the debate with defenders of the 'traditional orthodoxy' is far from over. For this reason I recommend the volume as supplementary reading material for advanced courses in linguistics, psychology, cognitive science and philosophy of language. The inclusion of facts about "numerous remote and little-known languages" (p.271), in particular, will make for an interesting addition to a subject matter that can sometimes be 'dry' to teach. I hope that many interested readers from various fields will find reading this volume worthwhile and that it will motivate further research into the intricate relationship between language structure, cognition and human culture.
Commissioned 5 Oct 2009
| Home | About ICLA | ICLA News | Events | Membership | Support the ICLA | Affiliates | Listservs | About Cognitive Linguistics | Study Cognitive Linguistics | Research and Teaching | Book reviews | Member homepages |
© 2002-present ICLA; all rights reserved.