Foundations of Language: Brain, Meaning, Grammar, Evolution

Citation

Jackendoff, R. (2002). Foundations of Language: Brain, Meaning, Grammar, Evolution. Oxford University Press.

Chapter Summary

The book is a single argument about what a theory of language should explain and how it should be organized. It has thirteen chapters in three parts, about 480 pages in all. Part I sets out the psychological and biological foundations that motivate generative linguistics, Part II presents the author's alternative architecture of the language faculty, and Part III works out a theory of meaning compatible with it. The author describes three concentric audiences: linguists of all persuasions at the centre; psycholinguists, neurolinguists, acquisition researchers and computational linguists around them; and, in the outer ring, psychologists, philosophers and others with a professional interest in language.

Part I, Psychological and Biological Foundations: Chapters 1–4

Chapter 1, "The Complexity of Linguistic Structure," uses the structure of a simple English sentence to show how much phonological, syntactic, semantic/conceptual and spatial structure is involved, and sets a two-part agenda: recognizing the levels and their tiers, and describing the correspondences between them. Chapter 2, "Language as a Mental Phenomenon," argues that language exists in the minds and brains of speakers, so that linguistics is a branch of psychology; it introduces the idea of a "functional mind" to avoid claiming a simple match between linguistic notation and neurons. Chapter 3, "Combinatoriality," concerns the unlimited number of sentences generated from finite resources, the kinds of rule involved (formation rules, derivational rules and constraints), and the problems combinatoriality poses for popular models of semantic memory and neural networks. Chapter 4, "Universal Grammar," clarifies common misconceptions about this term and places the author between a view in which linguistic structure is fully pre-specified and an empiricist view of acquisition.

Part II, Architectural Foundations: Chapters 5–8

Chapter 5, "The Parallel Architecture," is the theoretical core. It rejects the assumption that syntax is the only generative source of grammatical structure and proposes that phonology, syntax and semantics are independent generative systems, each with its own tiers, linked by interface components. Chapter 6, "Lexical Storage Versus Online Construction," discusses what is stored in long-term memory and what is built on line, and distinguishes productive from semi-productive morphology. Chapter 7, "Implications for Processing," reviews perception and production under the parallel model and argues for degrees of modularity rather than absolute modularity. Chapter 8, "An Evolutionary Perspective on the Architecture," asks how the language capacity could have evolved incrementally if the parallel model is correct, with later stages reflected in present-day language.

Part III, Semantic and Conceptual Foundations: Chapters 9–13

Chapter 9, "Semantics as a Mentalistic Enterprise," argues that meaning is an independent combinatorial structure and not reducible to syntax. Chapter 10, "Reference and Truth," considers how expressions relate to the world as conceptualized by the speaker, drawing on perceptual psychology. Chapter 11, "Lexical Semantics," concerns lexical decomposition, and Chapter 12, "Phrasal Semantics," develops a theory of how phrase meanings are composed, with parallels to lexical meaning. Chapter 13, "Concluding Remarks," draws the threads together.

Key Concepts

Chapters 1–4: Foundations

• Mentalism: the view that language is instantiated in speakers' minds and brains, making linguistics part of psychology.
• Combinatoriality: the capacity to produce and understand unboundedly many new sentences from a finite set of resources.
• Universal Grammar: the cognitive specialization for language, discussed against two extremes.

Chapters 5–8: Architecture

• Parallel architecture: independent phonological, syntactic and semantic generative systems connected by interfaces.
• Interface components: the links through which the generative systems correspond to each other.
• Lexical storage versus online construction: the distinction between stored items and structures built during use.
• Structure-constrained modularity: levels of structure communicate only through interfaces, giving degrees of modularity.

Chapters 9–13: Meaning

• Conceptual semantics: meaning as an independent combinatorial structure, linked to syntax and phonology by interfaces.
• Lexical decomposition: analysis of word meanings into components.
• Phrasal semantics: how phrase meanings are composed.

Critical Analysis

Strengths

• Interdisciplinary range. Reviewers stress the book's connections to psychology, neuroscience, biology and philosophy. Paloma Úbeda Mansilla, in Ibérica (2004), calls it a broad survey that is up to date and readable, and Pritha Chandra, writing on the LINGUIST List (2002), praises its effort to relate linguistics to other disciplines while examining existing generative proposals with scepticism.
• A single, clearly stated position. The three-part structure moves from the motivating questions, through the proposed architecture, to its consequences for meaning, so the reader can follow how each part depends on the previous one. Chapter 5 makes explicit which assumptions of standard generative theory the author rejects.
• Written for a wide readership. Úbeda Mansilla describes the style as readable and stylish and accessible to a wide readership, and Chandra notes that jargon is adapted for non-linguists.
• Evolution and semantics inside the theory. Evolution has its own chapter (Chapter 8) and meaning takes four (Chapters 9–12), so both are developed as parts of the argument rather than added at the end.

Limitations

• Advocacy for one model. The book presents the parallel architecture as an alternative to syntax-centred generative models rather than as a neutral survey; Úbeda Mansilla describes it as introducing the author's personal perspective on theoretical linguistics. It needs to be read alongside accounts of the approaches it criticizes.
• Assumes a general grounding. Although reviewers find the style accessible, the argument is defined against standard generative theory and moves among phonological, syntactic and semantic terminology, so it is easier for readers who already know the main areas of linguistics.
• Organized as an argument, not a survey. No chapter introduces phonology, morphology or syntax as such, so it cannot replace the broad textbooks for learning the subfields.
• Dated. It appeared in 2002, so later developments in the field are not covered.

Where It Fits in the Course

Jackendoff is not a first introduction to linguistics but a synthesis to read after one. Fromkin and colleagues' An Introduction to Language, O'Grady and colleagues' Contemporary Linguistics, Yule's The Study of Language and the modular Language Files textbook are comprehensive introductions for students new to the field. Radford and colleagues' Linguistics: An Introduction and Fromkin's edited Linguistics: An Introduction to Linguistic Theory go further into theory and analysis, with exercises, and are the better preparation for the technical parts of Jackendoff.

Read Jackendoff after one of the introductions, when the question has become how phonology, syntax and meaning fit together and how they connect to the mind, the brain and evolution. It contrasts usefully with the theory-driven survey of Fromkin and colleagues, which develops the analysis of each component in turn, and with introductions that treat syntax as the generative centre.

How to Use This Book

The book is for advanced undergraduates, graduate students and researchers in linguistics, cognitive science, psychology and philosophy of language who have already studied the main areas of linguistics. Some background in cognitive science helps.

A practical plan: read Chapters 1–4 for the framing, since they state the questions the rest of the book answers. Chapters 5–8 are the core and should be read in order, because the parallel architecture is introduced in Chapter 5 and developed in the next three. Part III can be read as a unit by those mainly interested in semantics, though it assumes the architecture of Part II. Chapter 8 can be read alone by readers interested in evolution, and Chapter 13 gives a short overview that is a good place to return to after the whole. Use the reading with an introduction to generative syntax to check the contrast with standard models.