Publishing Partner: Cambridge University Press CUP Extra Publisher Login

New from Cambridge University Press!


Revitalizing Endangered Languages

Edited by Justyna Olko & Julia Sallabank

Revitalizing Endangered Languages "This guidebook provides ideas and strategies, as well as some background, to help with the effective revitalization of endangered languages. It covers a broad scope of themes including effective planning, benefits, wellbeing, economic aspects, attitudes and ideologies."

New from Wiley!


We Have a New Site!

With the help of your donations we have been making good progress on designing and launching our new website! Check it out at!
***We are still in our beta stages for the new site--if you have any feedback, be sure to let us know at***

Review of  Grammar Without Grammaticality

Reviewer: Qizhong Chang
Book Title: Grammar Without Grammaticality
Book Author: Geoffrey Sampson Anna Babarczy
Publisher: De Gruyter Mouton
Linguistic Field(s): Linguistic Theories
Text/Corpus Linguistics
Issue Number: 25.4363

Discuss this Review
Help on Posting
Review's Editor: Helen Aristar-Dry


This book is a monograph comprising fifteen chapters explaining and giving evidence for a refreshing and novel view of the nature of human language, while questioning notions of grammaticality and empiricism that are widely accepted by most linguists. Other common and recurring themes are the authors’ arguments against the generative enterprise, rejecting the idea of using the linguist’s own intuitions as data, and warning against a Euro-centric view of language. The book draws extensively on data from several corpora and adopts a structural annotation scheme (SUSANNE) created by one of the authors (Sampson, 1995).

English provides most of the empirical material, and is both complemented and contrasted with other languages, such as Old Chinese, Malay and Indonesian. Most of the chapters in the book have been published as standalone papers previously, as part of an ongoing research program. The book is organized as follows:

1. Chapter 1: Introduction
2. Chapter 2: The bounds of grammatical refinement
3. Chapter 3: Where should annotation stop?
4. Chapter 4: Grammar without grammaticality
5. Chapter 5: Replies to our critics
6. Chapter 6: Grammatical description meets spontaneous speech
7. Chapter 7: Demographic correlates of speech complexity
8. Chapter 8: The structure of children’s writing
9. Chapter 9: Child writing and discourse organization
10. Chapter 10: Simple grammars and new grammars
11. Chapter 11: The case of the vanishing perfect
12. Chapter 12: Testing a metric for parse accuracy
13. Chapter 13: Linguistics empirical and unempirical
14. Chapter 14: William Gladstone as linguist
15. Chapter 15: Minds in uniform

In Chapter 1, the authors set out to develop the idea that languages have grammar, but not ‘grammaticality’. They show that an early grammarian, Meiklejohn’s (1902) failure to define a set of non-sentences (sentences that we consider ‘ungrammatical’) “...makes his description more faithful to the fundamental nature of human language than a theory that specifies a contrast between “grammatical” and “ungrammatical” word-sequences could be” (Sampson & Babarczy, 2013:5). The authors’ view grammar as a leaky system, stating that grammar is best described as “behavior patterns but no binding rules”. Sampson and Babarczy (S&B henceforth) claim that it is impossible to define grammaticality in the way logicians and generative grammarians attempt to do so, for there is no common logic that languages share -- it is not true that English and other human languages express a common range of thought-structures. To illustrate this, S&B show examples of vagueness in meaning in Indonesian, and Old Chinese. They also expect languages to continue to develop new thought-structures, with the accompanying grammatical innovation. Therefore, it would be pointless to talk of grammaticality.

S&B sets out to “enquire into the growth and limits of grammatical precision” (Sampson & Babarczy, 2013:24). To do this, they ask questions such as:

- If the grammatical structure of a language is developed by the community that uses it, and acquired by individual speakers, along lines that are not prescribed in advance, how refined does that structure become?

- Are there particular areas of grammar which are less, or more, precisely defined than other areas?

- Do the answers to some of these questions differ for written and spoken modes of language?

- Can we make generalizations about the path taken by children toward the levels of grammatical refinement achieved by adults?

According to S&B, the answers to these questions lie in the grammatical annotation of corpora data, instead of introspection and intuition.

In Chapter 2, the authors describe an experiment that tests the reliability of the grammatical annotation scheme SUSANNE: they compare the results when the scheme is applied independently to the same set of samples by separate analysts. Three conclusions fall out from the exercise: firstly, human error is more significant than definitional limitations; secondly, structural ambiguity is often pragmatically nonsignificant; and lastly, assigning functional categories is especially problematic.

Chapter 3 discusses various ways to approach indeterminacy in parsing structure, and the common/potential problems annotators face. S&B suggest that linguistic annotation ought not to be made dependent on linguistic theorizing, even when certain theories claim to have the answers to these problems of indeterminacy.

In Chapter 4, the authors discuss in detail their proposal and why they do not adhere to a simple notion of ‘(un)grammaticality’. They use the analogy that ungrammatical sentences are simply friends that one has not yet met. That is to say, certain sentences deemed to be ‘ungrammatical’ now might become widely used in the future. S&B objects to Chomsky’s (1957) claim that “the fundamental aim in the linguistic analysis of a language L is to separate the grammatical sequences...from the ungrammatical sequences...” and do not believe that there is a ‘range of things we cannot say’. The authors offer a piece of statistical analysis to prove their point: in a graph showing tagma frequency against proportion of tokens, they show that there is a significant number of distinct daughter-sequences which only appear once in a sample of 131, 302 words. In other words, constructions that are rarely observed, will together make up a large proportion of a language. These constructions are usually not found in grammatical descriptions of those languages. S&B also criticize the ‘unscientific’ idea that speakers of a language have introspective or intuitive access to aspects of its grammar: this could be idiosyncratic to each speaker, and there is no real reason to assume that patterns in a speaker’s intuitive grammaticality judgments reflect realities of his language. Arguments on empiricism and how intuition has led linguists astray are continued in Chapter 13. Chapter 4 ends with a discussion of the ‘realistic’ kind of descriptive (rather than prescriptive) grammar that linguists should be concerned with. The authors claim that the endeavor of trying to define ‘psychologically real’ grammars has not been very productive thus far.

In Chapter 5, S&B attempts to address known criticisms of their controversial proposal. These include objections from Pullum (2007) and Müller (2008). They also include support for their proposal from other scholars: for instance, Stefanowitsch (2007) believes that sequences commonly seen by linguists to be ‘ungrammatical’ often turn out to be absent from real-life usage not because of grammar, but because they mean things that people don’t want to say. The chapter, and also a large part of the remaining sections of the book, ultimately ends up being a critique of Universal Grammar. S&B claim that “...there are linguists for whom the concept of grammaticality is not an a priori assumption, but rather an implication which they need to believe in, because it follows as a consequence of a more abstract claim about human language to which they are attached” (Sampson & Babarczy, 2013:107). The authors also take issue with the fact that Universal Grammar has been poorly defended and documented in the linguistics literature, proponents not having provided many substantive and specific universals of language. S&B are also opposed to linguistic nativism, or Chomsky’s innateness hypothesis.

Chapter 6 gives us a good illustration of how spontaneous speech produces even more variant grammatical structures than written language, from an annotation point of view (using SUSANNE). This is an extension of S&B’s discussion of limited grammatical precision of spoken language. In Chapter 7, the authors examine demographic variables such as social class, gender and age, as well as speakers’ region of origin. Using data from the British National Corpus, they plotted mean embedding index against age, and obtained a steadily rising graph. To S&B, this shows a behavior of ‘lifelong learning’: that the process of language acquisition is a never-ending one; as opposed to the innatist belief that there is a critical period for language acquisition. The innateness hypothesis is further explored in Chapters 8 and 9, which discuss children’s writing and discourse. The authors provide evidence to show that there is NO innate programme under which grammar unfolds in a child’s mind in a fixed sequence.

In Chapter 10, S&B take issue with McWhorter’s claim that “the world’s simplest grammars are creole grammars” (McWhorter, 2001). Again, they use examples from Old Chinese and Malay/Indonesian, and Akkadian (from Deutscher, 2000) to show that there is no reason to expect all languages with long histories to contain any minimum structural complexity. They treat the development of novel logical or abstract concepts, and loss of existing categories, as normal aspects of the life of all human languages. Chapter 11 provides a case study of the vanishing Perfect distinction. This constitutes evidence that languages are constantly evolving, thus supporting the authors’ claims that there is no ‘grammaticality’.

The remaining chapters of the book compare different metrics for parse accuracy (Chapter 12), discuss empiricism in linguistics (Chapter 13), and discuss linguistic relativism, Eurocentrism, and linguistic diversity (Chapter 14 and 15).


An important merit of this book is that it asks linguists to confront their fundamental beliefs about language and approaches to research. The book bravely challenges the most established of linguistic theories -- the Generative framework -- and holds its own. The authors’ decision to devote a full chapter (Chapter 5) to answering their critics is a very welcome one, given the controversial nature of their proposal. However, some of their objections are too vague to be convincing. For instance, in their rejection of Meurers’s (2007) analogy of language as path-making in a mountainous area (where mountain peaks and other unpassable terrain represent ungrammaticality), the authors say “…we know that this is a logically-possible model; but we believe it to be the wrong model”, without pointing out explicitly what is wrong with Meurers’s model (Sampson & Babarczy, 2013:106). Some of the authors’ criticisms can also be construed as contradictory to their own position. In an example about the fallibility of using intuition as primary data, the authors criticize linguists who ‘invent’ data, by quoting Hanks (2013): “There is a huge difference between consulting one’s intuitions to explain data and consulting one’s intuitions to invent data. Every scientist engages in introspection to explain data. No reputable scientist (outside linguistics) invents data in order to explain it.” It could very well be the case, using the authors’ own analogy, that these ‘invented’ sentences are simply ‘unmet friends’.

The diversity of approaches adopted in this monograph -- from language parsing, annotation, language and mind, language acquisition, child language, to typology -- ensures that it will be relevant for linguists of any persuasion. For instance, Chapter 6 on the primacy of spontaneous speech to linguistic research, and Chapter 10 on how creole grammars are not simple grammars, covers themes explored in research on emerging varieties of English.

A significant number of the 15 chapters in this book come from papers previously published by one or both of the authors. Together, the majority of these chapters form a coherent, well-researched and detailed research program. However, at least two of these chapters look like they are ‘shoehorned’ into the book. Chapter 12, for instance, is a detailed comparison of the Leaf Ancestor metric and Grammar Evaluation Interest Group (GEIG) metric, which concludes in favor of the former. This comparison is probably too technical for the uninitiated and does not seem necessary for the authors’ proposal. The larger point made in this chapter is that most computational linguists accept the GEIG metric over the Leaf Ancestor metric, despite the authors’ claiming that the latter is superior. They believe that the only reason the GEIG metric is preferred over the Leaf Ancestor metric is that there is “authority” behind the former. This, at best, only lends indirect support to their overarching argument “…that the grammaticality concept has been accepted by linguists more generally: it has appeared to have authority behind it” (Sampson & Babarczy 2013:236); and in the first place, is a claim that cannot be objectively verified. I think the point could have been made more concisely. Chapter 14, an entire chapter about Gladstone’s writing about language, also seems odd and out of place; or at best, it is probably too long just to make the point that “intellectual advance requires not only individuals who produce good ideas but also an audience ready to receive them” (Sampson & Babarczy 2013:295).

This book is generally an excellent piece of writing, and suffers only from very infrequent spelling and formatting errors. Some chapters can be read independently of the others; but should be read together in sequence for a more coherent picture. The book is meant for advanced students and scholars of linguistics, who have experience in research, and a broad understanding of different subfields of linguistics, as well as a certain level of familiarity with the historical development of major theories within the field. The analyses are firmly grounded in a linguistic annotation framework, and cover much empirical ground. Overall, this is an important work that all linguists should read (even/especially if they are generative grammarians).


Chomsky, N. 1957. “Syntactic Structures”. Gravenhage: Mouton.

Deutscher, G. 2000. “Syntactic Change in Akkadian: the evolution of sentential complementation”. Oxford: Oxford University Press.

Hanks, P. 2013. “Lexical Analysis: norms and exploitations”. Cambridge, Mass.: MIT Press.

McWhorter, J. 2001. The world’s simplest grammars are creole grammars. “Linguistic Typology” 5: 125-166.

Meiklejohn, J. 1886. “The English Language: its grammar, history and literature”. 23rd ed., London: Alfred M. Holden.

Meurers, W. 2007. Advancing linguistics between the extremes. “Corpus Linguistics and Linguistic Theory” 3: 49-55.

Müller, S. 2008. “Head-Driven Phrase Structure Grammar: eine Einführung” (2nd revised edition). Tübingen: Stauffenburg Verlag.

Pullum, G. 2007. Ungrammaticality, rarity, and corpus use. “Corpus Linguistics and Linguistic Theory” 3: 33-47.

Sampson, G. 1995. “English for the Computer: the SUSANNE corpus and analytic scheme.” Oxford: Clarendon Press.

Sampson, G. & Babarczy, A. 2013. “Grammar Without Grammaticality: Growths and Limits of Grammatical Precision”. De Gruyter: Mouton.

Stefanowitsch, A. 2007. Linguistics beyond grammaticality. “Corpus Linguistics and Linguistic Theory” 3: 57-71.
Qizhong Chang is a PhD Candidate at the Department of Linguistics at the National University of Singapore. He is interested in Syntax, Syntax and new varieties of English.