LINGUIST List 31.2535

Mon Aug 10 2020

Review: Psycholinguistics: Adger (2019)

Editor for this issue: Jeremy Coburn <>

Date: 20-May-2020
From: María Florencia Silva <>
Subject: Language Unlimited
E-mail this message to a friend

Discuss this message

Book announced at

AUTHOR: David Adger
TITLE: Language Unlimited
SUBTITLE: The Science Behind Our Most Creative Power
PUBLISHER: Oxford University Press
YEAR: 2019

REVIEWER: María Florencia Silva, Universidad de Buenos Aires


‘Language Unlimited’ is a book that aims to answer what it is about human language that allows us to create and understand a limitless range of sentences. In that sense, Adger claims that humans are only able to use one pattern in syntax, which is hierarchical. Thus this approach is better than one that claims that we are powerful learners of patterns because of our high development of mental powers compared to animals. Furthermore, Adger argues that patterns that depend on sequences of words are invisible to us, while syntactic hierarchy is unavailable to other animals. In the ten chapters of ‘Language Unlimited’ Adger presents arguments and evidence to defend human language as something unique.

Chapter 1 (“Creating Language”): Adger outlines two perspectives for the study of human language. It’s possible to argue that either minds impose structure or humans learn different patterns from experience -a view that goes back to Darwin.

According to the first perspective, to answer how language allows us to create and understand a limitless range of sentences, three aspects should be tackled. The first one is organization: human languages are organized in a unique structurally way and our sense of that structure channels particular aspects of our linguistic experience into our minds.

The second aspect is that linguistic structures build meaning in a hierarchical way creating invisible relationships between words and thus these are governed by Laws of Language. Finally, the last one is that those laws that create the hierarchical structure of human language emerge through self-similarity (a property everywhere in Nature that makes a smaller structure echo the shape of a larger one). Because of this an unending collection of structures built from smaller structures is made available. Adger affirms that hierarchy and self similarity are what underlies our creative use of language.

Moreover, Adger points out that to explain the difference between (a) “The flea bit the woman” and (b) “The woman bit the flea” we can’t only rely on our mental dictionary to look for the meaning of these words and understand the meaning of the sentences; we also have to understand the specific structural combination that depends on what linguists call mental grammar.

Chapter 2 (“Beyond symbols and signals”): Adger raises the question about the possibility of building a universal language from iconic symbols.

Within the universe of symbols, there are iconic symbols, like emojis, with a direct link of resemblance between the meaning and the external form. However, we can also find arbitrary symbols associated with meanings through social conventions and “words are the crème de la crème of arbitrary symbols” (23).

Furthermore, Adger pinpoints that emoji language couldn’t be universal and, in the absence of context, they are too vague to work like language. He asks, for example, how we would express an event in the past using just emojis.

The author finds human language goes beyond symbols because it has ways and means of expressing different ideas. If we impose a meaning convention for a sequence in emoji language (first the individual who performs, then the action, and finally the individual affected by the action), how would we explain sentences in which the first individual is the one affected by the action, and not the one who performs it?

In addition, one of the most important properties of sentences of human languages is that words cluster together in groups, and languages are sensitive to this. If we read the following sentences (a) “[The person who can run fastest] can win the marathon”, (b) “[The person who can catch the cat that can run fastest] can win the marathon” and (c) “[He] can win the marathon”, we can observe that “he” works as a perfect replacement because the words between brackets in (a) and (b) behave as a single group.

Finally, the chapter ends with a classic question: “is language just communication?” (35). Adger states that, although language is used to communicate or exchange information, this doesn’t mean that the nature of language has to be communicational (a lot of things communicate). Language is used for many other things and some of them are a side effect of its use as a tool for communication. Furthermore, we also use language to express, to order, and create our thoughts.

Adger claims “communication is one of the things language can be used for, certainly, but just because something is used to do something, that doesn’t tell us what that something is” (39). We need to understand the structure that underlies language.

Chapter 3 (“A sense of structure”): Adger says it’s well known that the way in which words can be put together affects interpretation. This fact can trigger syntactic ambiguity, a property related to structure; e.g., if we read the sentence “public park or playground”, “public” can modify only the first noun or both. Adger claims, following Jespersen (1924), that “human beings appear to have an ability to unconsciously sense what the structure is when we hear them” (46) and that this ability also guides us when we produce sentences.

Furthermore, if we want to understand when the notion of structure appears in our life, we have to take into consideration that when we are babies we have to figure out what variety of human language we are immersed in and in order to do that we are provided with an innate resource: Universal Grammar (Chomsky 1986). The author outlines that UG (Laws of Language), plus linguistic external experience and general intellectual skills allow us to develop the notion of structure of our own language.

Chapter 4 (“The question of Psammetichus”): this chapter deals with the two perspectives on language mentioned before (Chapter 1): either our sense of structure comes from our biological setup or it is the result of general cleverness and social interaction. The latter understands experience as the material from which we learn, while the former understands the structure of language as part of our nature as humans and the ability to recognize structure as a special one.

According to this view, learning a language is a process where we subconsciously figure out which structures work out best to help both understand what we heard and express what we want to say.

Adger presents one piece of evidence in defense of this view: homesign language (Goldin-Meadow 2005). Homesign language is a system of gestures developed by deaf children whose parents are hearing (so they don’t have sign language at all). Homesigners create a language on their own without any linguistic input.

This language has properties of other spoken or signed languages; for example, homesign clusters sign together into units or recreate the distinction between nouns and verbs and display abstract syntactic structure. So far so good, if we acquired language only by extracting patterns present in linguistic experience, homesign language couldn’t exist. Adger claims we can produce and understand structure because of our innate biological ability.

Chapter 5 (“Impossible patterns”): Adger states that what languages have in common is structure: particular building blocks; and that those parts of the structure are organized by the rules of language. What distinguishes a language from random sounds is the underlying structure in the former and our ability to sense that structure at play. Furthermore, grammatical concepts (Number, Tense, Subject and Object) are linked with discrete categories, they crop up across languages, and categorize and alter words so they can be combined into sentences.

Chapter 6 (“All in the mind”): categorial perception is the phenomenon that explains why we don’t perceive gradualness; since our language works with discrete categories, we impose a boundary between two classes rather than a continuum. Humans have an inbuilt bias to turn something apparently continuous into discrete categories (an ability we share with other animals) and to pay attention to certain properties of sound (e.g., humans can extract wordlike units from streams of syllables by paying attention to how likely certain consonants are to follow others in the stream of sound, but they can’t do this by paying attention to the vowels and ignoring consonants -while tamarin monkeys can).

Adger pinpoints that although the capacity to conceive the world as involving discrete units and to connect sounds and concepts is shared with animals, this connection can be done over tens of thousands times by humans, whereas other animals are limited to a few tens of calls. Finally, the author outlines that “some of the general abilities that we humans have, that enter intimately into our capacity to use language, are shared with other species of animals, but the details of how those abilities work may differ from animal to animal” (141).

Chapter 7 (“A law of language”): the chapter starts with the well-known story of Kanzi, a bonobo whose abilities with language were claimed to be comparable in some way to human language (Savage-Rumbaugh 1986). However, after a few years, linguist Rob Truswell showed in his study (2017) that Kanzi lacked sense of hierarchy, since it was using order, and not structure. When Kanzi was asked to react to sentences with a coordinated structure like “Give the water and the doggie to Rosie” or “Give me the milk and the lighter”, the bonobo ignored the first element, or the second, or took both. This kind of chunking conjoins the various words (“water”, “and”, “doggie” or “milk”, “and”, “lighter”) into a larger unit ([water and doggie] or [milk and lighter]) creating a hierarchical structure for the sentence.

Adger raises the question about how we learn hierarchy if syntactic structure isn’t an audible part of sentences.

As it was pointed out in chapter 3, we use linguistic experience and Laws of Language (UG), which connect and constrain what we mean and what we say, to create a system of rules known as structure. Arguments in defense of this view come from homesign language (Chapter 4), the usage of different parts of the brain to process possible versus impossible languages (Chapter 5), and the way in which patterns of language work in English (e.g., coreferentiality between a pronoun and a name).

Chapter 8 (“A botlang”): A debate about new types of AI and their understanding of human language is introduced. He emphasizes that AI assistants don’t understand sentences like us, rather they learn a language whose rules are based on visible sequences, not invisible hierarchical structures like humans do.

Chapter 9 (“Merge”): Adger claims self-similarity (Chapter 1) is at the heart of syntax and Merge, as a recursive process (it reuses its own output), is an example of this principle: “it takes two bits of language [...] and creates out of them another bit of language” (211) and so on, thus creating unlimited language. The output of Merge is not a sequence but a new hierarchical unit.

So far so good, language builds up meaning through Merge and “each Merge comes along with an effect on the meaning of the sentence, and that effect is both stable, and quite systematic, no matter what language [we] are looking at” (215).

Chapter 10 (“Grammar and culture”): the author raises the question about how language is used to express our social selves. He pinpoints that there’s a difference between grammatical (past, plural, and so on) and social (identity, affiliation, class) meaning. The former is linked with discrete categories (Chapter 5) that Merge manipulates, while the latter can be linked to continuous aspects of languages; e.g., using of vowels as a belonging mark to some social group (Eckert 2000).

When we use language, we are creating sentences through Merge -that’s linguistic structure- but we also choose words and make different sounds to express our identities, our emotions, and our intentions.

It’s evident that a division of labour in language exists, but “both linguistic and social meaning are fundamental parts of the system we use in our day to day lives” (237).


It’s always difficult to write a book about linguistics for a general audience and that’s the reason why the most important merit of ‘Language Unlimited’ is its clear way to expose arguments in defense of the uniqueness of human language. Adger’s goal of writing a book not “about how to do linguistics [but a book] about what the nature of language is” (173) was achieved through a pedagogical way of introducing theoretical concepts like Merge or Core Knowledge, as well as evidence in defense of his arguments.

‘Language Unlimited’ gives a detailed and precise account of different topics related to language and answers some common questions such as “is language just communication?” (Chapter 2) or “do animals have language like us?” (Chapter 7). Furthermore, the introduction of issues such as the using of emoji language brings novelty to the book and makes it of interest non-expert and young audiences. Other issues that ‘Language Unlimited’ explores, include various topics, such as animal and AI communication and constructed languages (in Chapter 5, Adger tells how he created Waring, a language for the Beowulf TV series), that address common questions, such as the ones pointed out above.

One of the highlights of the book is the evidence against those perspectives which argue in favour of learning language from experience: the description of homesign language and all the examples presented in Chapter 4, not only allow Adger to defend our biological setup but also one of the main topics of the book: the issue of structural hierarchy in language. Furthermore, the hierarchical issue comes up with the introduction of Merge, an argued concept in linguistic literature, that despite Adger’s clear explanations, opens up a lot of questions known for linguists but not for the general audience.

Nevertheless, ‘Language Unlimited’ is an impressive book thanks to the excellent exposition David Adger makes and, above all, because of the addition and precise discussion of examples (e.g., in Hixkaryana, in English, in Nuuchanulth, and in German, among other languages) and theories (e.g., Chunking approach). Everything said, ‘Language Unlimited’ is a must for any linguistic enthusiastic.


Chomsky, Noam. 1986. Knowledge of Language: Its Nature, Origin, and Use. New York: Praeger.

Eckert, Penelope. 2000. Linguistic Variation as Social Practice. Oxford: Blackwell.

Goldin-Meadow, Susan. 2005. The Resilience of Language. New York: Psychology Press.

Jespersen, Otto. 1924. Philosophy of Grammar. London: Allen & Unwin.

Savage-Rumbaugh, Susan. 1986. Ape Language: From Conditioned Response to Symbol. New York: Columbia University Press.

Truswell, Robert. 2017. Dendrophobia in bonobo comprehension of spoken English. Mind and Language 32: 395–415.


María Florencia Silva is BA, Major in Linguistics (University of Buenos Aires). Her main topic of research is syntax, in particular issues related to head movement. As part of her research in syntax, she has previously written an article about the interaction between grammatical aspect and adverbs (“The interaction between grammatical aspect and adverbs”).
She is also interested in the semantic information of mental representations and issues in philosophy of linguistics. Currently, she is a member of the Buenos Aires Philosophy of Cognitive Science Research Group and works as transcriber in the research project ‘Effects of variation and variability in the acquisition of two dialects of Spanish’ (Michigan State University).

Page Updated: 10-Aug-2020