The innate grammar blueprint: What is it, and why isn’t it a crazy idea?

Many linguists assume that languages are made up of the same basic building blocks – not the same words, of course, but the same phonological features (e.g. Chomsky & Halle 1968), the same morphosyntactic features (e.g. Corbett 2012), the same semantic primitives (e.g. Goddard (ed.) 2008), the same types of rules or constraints, and the same components (e.g. syntax vs. morphology), as well as the same overall architecture (e.g. Jackendoff 1997).

This set of basic building blocks has been widely thought to be innately given, as part of the linguistic genotype, i.e. the genetic basis of human linguisticality (the capacity for language, Haspelmath 2020). They are thought to be part of human nature, and can be thought of as natural kinds (Haspelmath 2018). The entire set of pre-established building blocks which languages are made up of has been aptly called the grammar blueprint. A very clear formulation of this idea, from the perspectve of language acquisition, is provided by Pinker (2007):

“In cracking the code of language…children’s minds must be constrained to pick out just the right kinds of generalizations from the speech around them….It is this line of reasoning that led Noam Chomsky to propose that language acquisition in children is the key to understanding the nature of language, and that children must be equipped with an innate Universal Grammar: a set of plans for the grammatical machinery that powers all human languages. This idea sounds more controversial than it is (or at least more controversial than it should be) because the logic of induction mandates that children make some assumptions about how language works in order for them to succeed at learning a language at all. The only real controversy is what these assumptions consist of: a blueprint for a specific kind of rule system, a set of abstract principles, or a mechanism for finding simple patterns (which might also be used in learning things other than language).” (Pinker 2007)

Notice that Pinker also uses the better-known term “universal grammar” (UG), but he says that there are three ways in which UG may be understood: (i) as a grammar blueprint, (ii) as “a set of abstract principles”, or (iii) as a domain-general mechanism. I will say a bit more about “UG” below, and here I will focus on the more specific notion of a grammar blueprint.

The ingredients of the grammar blueprint are often said to be “formal universals” and “substantive universals”:

“It is useful to classify linguistic universals as formal or sub­stantive. A theory of substantive universals claims that items of a particular kind in any language must be drawn from a fixed class of items…. A theory of substantive semantic universals might hold, for ex­ample, … that each language will contain terms that designate persons or lexical items referring to certain specific kinds of objects, feelings, be­havior, and so on. …. The property of having a grammar meeting a certain abstract condition might be called a formal linguistic universal… For example, consider the proposal that … the phonological component of a grammar consists of a sequence of rules, a subset of which may apply cyclically to successively more dominant con­ stituents of the surface structure…” (Chomsky 1965: 28-29)

In practice, this has been a very influential (if not dominant) view in general grammar research since the 1970s, and it continues to be assumed widely. For example, textbooks present their readers with sets of phonological features (e.g. Hayes 2009: §4.3), sets of syntactic features and rule types (e.g. Sportiche et al. 2014), and so on.

Another metaphor that has sometimes been used is that of a toolkit, which is basically the same as a blueprint:

“Universal Grammar is not supposed to be what is universal among languages: it is supposed to be the “toolkit” that a human child brings to learning any of the languages of the world. If we find that a certain aspect of linguistic structure is indeed universal, then it is a good candidate for part of Universal Grammar…  However, non-universal aspects of linguistic structure may be candidates for Universal Grammar as well. When you have a toolkit, you are not obliged to use every tool for every job. Thus we might expect that not every grammatical mechanism provided by Universal Grammar appears in every language. For instance, some languages make heavy use of case marking, and others don’t; some languages make heavy use of fixed word order, and others don’t.” (Jackendoff 2002: 75)

So the idea is that the kinds of concepts that structural linguists use in their everyday work – consonant, syllable, noun, affix, relative clause, morphology, quantifier – are part of this blueprint or toolkit. Here are a two more quotations that show the use of the word “blueprint”:

“Children are preprogrammed to adhere to these principles of linguistic analysis as part of the blueprint for their development.” (Crain & Thornton 1998: §3.6)

“UG is proposed as part of an innate biologically endowed language faculty… UG provides a genetic blueprint, determining in advance what grammars can (and cannot) be like.” (White 2003: 2)

The term is also used by critics of the blueprint idea, e.g.

“Syntactic representations in this type of efficiency-driven computational system have the familiar binary-branching design – but not as the result of am a priori grammatical blueprint such as the X-bar schema.” (O’Grady 2003: 47)

“in the latter half of the 20th century, Noam Chomsky took on the challenge of understanding and investigating this blueprint, by looking to biology as the source of grammar, proposing that all languages are simply local manifestations of a biologically transmitted Universal Grammar. Other researchers believe that grammar is in essence a local phenomenon. This is the alternative I explore in what follows. In this view language and its components (grammar, phonetics, phonology, semantics, and so on) are perceived as local, cultural-communicational outputs, with little or no evidence for a genetic blueprint for grammar. “ (Everett 2016)

“In this [= the generative] framework, a child’s language ability gradually unfold according to a genetic blueprint in much the same way as a chicken grows a wing (Chomsky 1988: 150-151).” (Christiansen & Chater 2016: 21)

There is, of course, much uncertainty around the idea that many aspects of grammar are so similar across languages because they are innately pre-specified. So although it is a very widespread idea, it is not by any means firmly established, and in 2007, I even wrote a paper with the title “Pre-established categories don’t exist”. But the title was provocative, and the main thrust of the paper is summarized by its subtitle: “Consequences for language description and typology”. Since there is so much uncertainty about the blueprint and its building blocks, we cannot presuppose them in our everyday work, either for language description or for cross-linguistic comparison. (Comparison must be based on measurement uniformity, not on building-block uniformity.)

However, innate building blocks that form a grammar blueprint are not a crazy idea. I noted this in my 2020 paper on linguisticality:

“Since there is no doubt about the biological basis of human linguisticality, it is perfectly possible that not only the instinct to communicate, to imitate and to extract patterns from observed speech signals is innate, but that also a substantial number of specific structural building blocks (features, categories, and architectures) are in place before children start hearing their caretakers speak. The capacity for language would be like the capacity for taste, where culture-specific taste categories (which enable culture-specific recipes and cuisines to exist and to be transmitted) coexist with (and have an ultimate basis in) five innate basic taste categories (sweet, sour, salty, bitter, umami). (Haspelmath 2020: 3)”

The discovery of a basic set of taste categories was very surprising, given that the range of diverse tastes that we experience is extremely broad. Thus, it is very sensible to ask whether the apparent diversity of linguistic patterns in the world’s languages reduces to a smallish set of innate elements. I’m still not very optimistic about the successes so far, but the programme is perfectly reasonable. (What does not seem reasonable, however, is to use the same building blocks for all languages and to deny that they are innate, as is sometimes done. I have written more about this in my forthcoming paper “General linguistics must be based on universals”.)

Finally, let us briefly ask: How does the term “universal grammar” relate to the notion of an innate grammar blueprint? This is difficult to say, because the expression “UG” has been used in a variety of different senses (cf. Haspelmath 2020: 3): UG can mean (1) the same as innate grammar blueprint (as in the quotation from Jackendoff above); (2) the domain-specific aspects of human linguisticality, so that we can talk about a “UG hypothesis” which may or may not be right; and (3) UG can simply stand for “the striking difference in cognitive capacity between [humans and non-human animals] “ (Chomsky et al. 2019: 230). Colloquially, “UG” is most often used in the older sense (1), but it is my impression that this usage is getting rarer, as there is less and less confidence in the idea of a grammar blueprint.

In any event, I find the term innate grammar blueprint fairly clear, so I now prefer it to other terms that are much less clear (especially “universal grammar”, and “faculty of language (in the narrow/broad sense)”).

Before concluding, I should briefly discuss two recent commentaries on my “General linguistics” paper. First, José-Luis Mendívil-Giró writes:

“[G]enerative grammarians do not assume that there is a “rich innate grammar blueprint” and that languages are composed of “innate building blocks of grammar”. What they assume is that, if the capacity of language is innate …, then it would be very surprising that this capacity did not channel, condition or restrict the design space available to children, who have to acquire the language of their environment.” (Mendívil-Giró 2021)

Mendívil-Giró is right about the “official” position of many generative grammarians who have followed Chomsky’s “minimalist” turn, away from the substantive universals that he assumed in 1965 (see the above quotation), toward a vision in which very little is thought to be innate. But it is not clear to me that this is more than a vision, because generative grammar papers and textbooks typically make use of universal building blocks in a way that one would not do if one were to adopt an agnostic attitude. Mendívil-Giró’s commentary does not go into any details, but it seems to me that he himself (e.g. in his 2020 paper on inflection and derivation) uses terms like “affix”, “root”, “word”, “lexeme”, “phrase” in a way that presupposes their innateness, as part of a universal grammar blueprint.

In any event, to the extent that grammatical theorists want to explain limits on cross-linguistic variation, they must assume that their framework is innate, i.e. they must conceive of it as a grammar blueprint, rather than merely as a notational system. If a formal framework is not restrictive, then it is not explanatory but merely serves as a notation (see also Haspelmath 2014 on restrictivist vs. non-aprioristic approaches). In her commentary on my “General linguistics” paper, D’Alessandro (2021) says about her generative approach:

“We are not asking what is there, but what can be. A theory makes predictions not only as to what will happen, but also, and especially, as to what will not.”

But the generative approach does not consist of sets of predictive statements (like Greenberg’s (1963) 45 universals). It consists of a set of architectures, categories and features, and if these are not thought to be innate, then no predictions are made. Since I am interested in cross-linguistic predictions, I find views in the Chomskyan tradition that reject the grammar blueprint much less interesting than the proposals of the earlier the 20th century approach (that is epitomized by works like Roberts 1996; Baker 2001).

References

Baker, Mark C. 2001. The atoms of language. New York: Basic Books.

Chomsky, Noam & Gallego, Ángel J. & Ott, Dennis. 2019. Generative grammar and the faculty of language: Insights, questions, and challenges. Catalan Journal of Linguistics 2019. 229–261. (doi:10.5565/rev/catjl.288)

Chomsky, Noam & Halle, Morris. 1968. The sound pattern of English. New York: Harper & Row.

Chomsky, Noam A. 1965. Aspects of the theory of syntax. Cambridge, MA: MIT Press.

Chomsky, Noam A. 1988. Language and problems of knowledge: The Managua lectures. Cambridge, MA: MIT Press.

Christiansen, Morten H & Chater, Nick. 2016. Creating language: Integrating evolution, acquisition, and processing. Cambridge MA: MIT Press.

Corbett, Greville G. 2012. Features. Cambridge: Cambridge University Press.

Crain, Stephen & Thornton, Rosalind. 1998. Investigations in universal grammar. Cambridge, MA: MIT Press.

D’Alessandro, Roberta. 2021. Not everything is a theory: A reply to Martin Haspelmath. Theoretical Linguistics (to appear). (https://ling.auf.net/lingbuzz/005659)

Everett, Daniel L. 2016. An Evaluation of Universal Grammar and the Phonological Mind1. Frontiers in Psychology. Frontiers 7. (doi:10.3389/fpsyg.2016.00015)

Goddard, Cliff (ed.). 2008. Cross-linguistic semantics. slcs.102. Amsterdam: Benjamins. (https://benjamins.com/catalog/slcs.102) (Accessed February 24, 2021.)

Haspelmath, Martin. 2007. Pre-established categories don’t exist: Consequences for language description and typology. Linguistic Typology 11(1). 119–132.

Haspelmath, Martin. 2014. Comparative syntax. In Carnie, Andrew & Sato, Yosuke & Siddiqi, Dan (eds.), The Routledge handbook of syntax, 490–508. London: Routledge. (https://zenodo.org/record/344909)

Haspelmath, Martin. 2018. How comparative concepts and descriptive linguistic categories are different. In Van Olmen, Daniël & Mortelmans, Tanja & Brisard, Frank (eds.), Aspects of linguistic variation: Studies in honor of Johan van der Auwera, 83–113. Berlin: De Gruyter Mouton. (https://zenodo.org/record/3519206)

Haspelmath, Martin. 2020. Human linguisticality and the building blocks of languages. Frontiers in Psychology 10(3056). 1–10. (doi:10.3389/fpsyg.2019.03056)

Hayes, Bruce. 2009. Introductory phonology. Malden, MA: Wiley-Blackwell.

Jackendoff, Ray. 2002. Foundations of language: Brain, meaning, grammar, evolution. Oxford: Oxford University Press.

Jackendoff, Ray S. 1997. The architecture of the language faculty (Linguistic Inquiry Monographs 28). Cambridge, MA: MIT Press.

Mendívil-Giró, José-Luis. 2021. On the innate building blocks of language and scientific explanation: A reply to Haspelmath. Theoretical Linguistics (to appear). (https://ling.auf.net/lingbuzz/005708)

O’Grady, William. 2003. The radical middle: Nativism without universal grammar. The Handbook of Second Language Acquisition, 43–62. Malden, MA: Wiley. (doi:10.1002/9780470756492.ch3)

Pinker, Steven. 2007. The stuff of thought: Language as a window into human nature. New York: Viking.

Roberts, Ian. 1996. Comparative syntax. London: Edward Arnold.

Sportiche, Dominique & Koopman, Hilda Judith & Stabler, Edward P. 2014. An introduction to syntactic analysis and theory. Chichester: Wiley.

White, Lydia. 2003. Second language acquisition and universal grammar. Cambridge: Cambridge University Press.


One thought on “The innate grammar blueprint: What is it, and why isn’t it a crazy idea?

  1. Twitter comment from David Adger: “I think you’ve misunderstood Pinker’s notion of blueprint here. He was distinguishing between the rule-format idea of UG and the Principles/Parameters idea. The former says you learn rules but they have to fit into a certain format, and was basically rejected in the early 80s. The latter that there are no rules, just interacting principles. So the sense of a blueprint Pinker was getting at is tied to that rule-format notion of UG (basically Aspects onwards), and it’s definitely not the Baker/Rovberts etc approach, which is firmly Principles Parameters.” https://twitter.com/davidadger/status/1364571533994123269

    Haspelmath‘s reply: What I need is a cover term for “Aspects” and “Principles & Parameters”. Chomsky presented this as a big shift, but he kept the idea that features/categories and architectures are given in advance (= are innate). So I subsume both under “20th century Chomskyan”.

    Adger: But Minimalism still does this, it just says that you can break things down into more abstract elements that do more theoretical work, and are innate, and recruited for language. So conceptually, I think there is far less of a shift from P&P to Minimalism than from Aspect to P&P

    Haspelmath: Tecumseh Fitch seems to see a big shift: “In another revolutionary reinvention of linguistics, C has reconceptualized the core of language as consisting of a single very powerful but abstract capacity that is unique to humans.” (https://inference-review.com/article/reinventing-linguistics-again)

    Adger: I think that’s wrong though (and I think Chomsky would think it’s wrong too). It’s not a change in the conception of what a grammar is (which P&P was), rather a change in what it contains (less than we thought).

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.