The Architecture of the Language Faculty
Ray Jackendoff, 1997

Jackendoff initially states that this book is a rehash of his 1987 Consciousness and the Computational Mind, but then proceeds to list a huge number of modifications and alterations since that book.

Chapter 1. Questions, Goals, Assumptions
"The present study is an attempt to renovate the foundations of linguistic theory".

1.1 Universal Grammar
While Jackendoff asserts that this book is "down to its deepest core a part of the Chomskian tradition", he also states that this involves "the way a human being understands and uses language". Those goals have not always been a part of Generative Linguistics. This book will not be concerned with social aspects of language use and it will generally assume that discourse and narrative issues are independent of sentential issues. It will not be overly formal. It is clear that children need something beyond general intelligence to acquire language. This extra something we call Universal Grammar. That this UG must be innate has made waves across many disciplines. Are UG "rules" represented in the brain a)not at all, b) indirectly, as interpretive data, or c) directly as procedural structure?

1.2 Necessities and Assumptions
Chomsky identifies three interfaces the grammatical computational system must have. These are 1) articulatory and perceptual systems, 2) the conceptual/intentional system, and 3) the lexicon. These interfaces are considered to be conceptual necessities of the theory. Among a number of additional assumptions, he includes that the system performs derivations, that it can substitute one string for another, that lexical items are combined by merging to build up phrase structures, that lexical items are inserted into the phrase structures, and that the grammar is nonredundant, entirely optimally coded.

1.3 Syntactocentrism and Perfection
Two further Chomskian assumptions present greater problems for Jackendoff. One, that syntax is the fundamental component of the system, Jackendoff refers to as syntactocentrism. He takes this opportunity to review a number of problems with several of the assumptions of Chomskian theory and, in particular, very cautiously introduces his view that the study of language must, to some extent, concern itself with the real workings of language rather merely describing the observable structures. He sums up a final assumption, that language approximates a perfect system, by his interpretation of a Chomskian view that "language could be perfect if only we didn't have to talk".

Chapter 2. Interfaces; Representational Modularity
This chapter looks more closely at the articulator/perceptual and conceptual interfaces, aguing that they play larger roles in the language system than the Chomskian view would support. Jackendoff then presents his conception of Representational Modularity.

2.1 The "Articulatory-Perceptual" Interfaces

2.2 The Phonology-Syntax Interface

2.3 The Conceptual-Intentional Interface

2.4 Embedding Mismatches between Syntactic Structure and Conceptual Structure

2.5 The Tripartite Parallel Interface

2.6 Representational Modularity

Chapter 3. More on the Syntax-Semantics Interface

3.1 Enriched Composition
Jackendoff argues that

3.2 Aspectual Coercions
In the sentence "The light flashed until dawn", an interrupted temporal sequence is understood which is not contained in the semantic information of any of the lexical items. This temporal structure must be "coerced" from the combination of items; that "until", as a temporal delimitor, forces an interpretation of repeated sequence onto the otherwise singular action of flashing.

There is a similar coercion in conversions between mass and count nouns imposed by specific pragmatic situations. However, these are restricted to cases of food or drink and do not appear to generalize beyond that.

3.3 Reference Transfer Functions
Several cases are presented in which an otherwise unremarkable NP is expanded so as to refer to a different pragmatically understood object.

Each of these three types of semantic modification are taken as examples of the semantic influence of .. (enrichment), which Chomsky argues must be semantically isolated.

This last comment is not well stated. Clean it up.

3.4 Argument Structure Alternations

3.5 Adjective-Noun Modification

3.6 Summary

3.7 Anaphora

3.8 Quantification

3.9 Remarks

Chapter 4. The Lexical Interface

4.1 Lexical Insertion versus Lexical Licensing

4.2 PIL = CIL

4.3 PIL and CIL Are at S-Structure

4.4 Checking Argument Structure

4.5 Remarks on Processing

4.6 The Lexicon in a More General Mental Ecology

Chapter 5. Lexical Entries, Lexical Rules

5.1 Broadening the Conception of the Lexicon

5.2 Morphosyntax versus Morphophonology

5.3 Inflexional versus Derivational Morphology

5.4 Productivity versus Semiproductivity

5.5 Psycholinguistic Considerations

5.6 "Optimal Coding" of Semiproductive Forms

5.7 Final Remarks

Chapter 6. Remarks on Productive Morphology

6.1 Introduction

6.2 The Place of Traditional Morphophonology

6.3 Phonological and Class-Based Allomorphy

6.4 Suppletion of Composed Forms by Irregulars

6.5 The Status of Zero Inflections

6.6 Why the Lexicon Cannot Be Minimalist

Chapter 7. Idioms and Other Fixed Expressions

7.1 Review of the Issues

7.2 The Wheel of Fortune Corpus: If It Isn't Lexical, What Is It?

7.3 Lexical Insertion of Idioms as X0s

7.4 Lexical Licensing of Units Larger Than X0

7.5 Parallels between Idioms and Compounds

7.6 Syntactic Mobility of (Only) Some Idioms

7.7 Idioms That Are Specializations of Other Idiomatic Constructions

7.8 Relation to Construction Grammar

7.9 Summary

Chapter 8. Epilogue: How Language Helps Us Think

8.1 Introduction
Jackendoff does not believe that language is thought, but rather that thought is a brain function separate from language, and which can proceed without language. Language opens the door to additional forms of more complex reasoning than are possible without it.

8.2 Brain Phenomena Opaque To Awareness
We are unaware of most sensory processing, becoming aware of only the end result. Consciousness is this awareness of certain sensory processing results. But which ones and what are the consequences?

8.3 Language Is Not Thought and Vice Versa
The fact that we can translate a thought between different languages shows the the thought exists separate from language. Various types of brain malfunctions highlight this distinction. The assignment of meaningful elements (theta roles) must be a conceptual issue, not a language issue.

8.4 Phonetic Form Is Conscious, Thought Is Not

8.5 The Significance of Consciousness Again

8.6 First Way Language Helps Us Think: Linguistic Communication

8.7 Second Way Language Helps Us Think: Making Conceptual Structure Available for Attention

8.8 Third Way Language Helps Us Think: Valuation of Conscious Percepts

8.9 Summing Up

8.10 The Illusion That Language Is Thought

Appendix. The Wheel of Fortune Corpus

Top of Page | ALF Opinion | Sort by Topic | Sort by Title | Sort by Author