This book presents a new theory of grammatical categories - the Universal Spine Hypothesis - and reinforces generative notions of Universal Grammar while accommodating insights from linguistic typology.
EDITORS: Rochelle Lieber and Pavol Štekauer TITLE: The Oxford Handbook of Compounding PUBLISHER: Oxford University Press YEAR: 2009
Fredrik Heinat, Stockholm University
The Oxford Handbook of Compounding has the intention of being a resource for scholars working on compounds and also a starting point for the non-initiated. The book contains a short presentation of the contributors, 34 chapters divided in 2 parts, references and an index. The first part deals with compounds from a theoretical point of view. The 16 chapters in this part concentrate on how compounds are treated in various theoretical frameworks, from a synchronic, diachronic, psycholinguistic and developmental perspective. The second part has a typological approach and contains chapters on 17 different languages from various language families. This is, according to the editors, the largest number of languages compiled on the topic.
Chapter 1: Introduction: status and definition of compounding, by Rochelle Lieber and Pavol Štekauer, pp. 3-18. In the introductory chapter the editors start by discussing definitions of compounds. They identify two problems that make it difficult to come up with a general definition of compounds that holds crosslinguistically; first, in some languages, the elements in a compound are not free standing words (hence the definition cannot be that a compound is a combination of two or more words) and second, in some languages it is not clear how to make a distinction between compounds and derived words. The following sections look at phonology, syntax, and inflection and the editors' conclusion is that in none of these areas can we find criteria for picking out compounds and that maybe compoundhood is best seen as a scale from more to less compoundlike. The chapter concludes with an outline of the book. The editors also distinguish what seem to be the three main problems that crop up in most of the chapters: the definitional problem, the problem of interpretation, and the component problem i.e. what do compounds tell us about the 'architecture of grammar' and the relationship and interaction of its parts (syntax, semantics, and morphology).
Chapter 2: Compounding and idiomatology, by Stanislav Kavka, pp. 19-33. Kavka claims that compounds share features and properties with idiomatic expressions, e.g. 'the way they arise, their existence proper and their interpretation' (p. 32). He says that compounds, just like idioms, are 'highly conventionalized, context bound expressions'. The aim of the chapter is to 'give evidence by citing some prototypical examples' (p. 28). Productivity is explicitly something that is NOT going to be explained. Of course some compounds are like idiomatic expressions, but the author picks a few of those and disregards the productive processes of compounding. Kavka also disregards the fact that there are several types of compounds and that the processes governing their semantic interpretation may be of different kinds.
Chapter 3: The classification of compounds, by Sergio Scalise and Antonietta Bisetto, pp. 34-53. Scalise and Bisetto go through some problems that arise in the classification of compounds. Terminological problems arise because researchers choose to include or exclude certain types of compounds under the same term. Another problem is narrow empirical coverage. Compounds of other types than N-N and Adj-N, such as Adj-Adj, V-V and others, have received surprisingly little attention in the categorization of compounds. A final problem the authors point out is the fact that researchers have used different criteria in categorizing compounds. Particularly the difference between endo- and exocentric compounds and the presence or lack of derivational morphology seem to be the cases where proposals differ the most. In the final sections they review one of their previous categorizations and propose a new one. They suggest that compounds can be divided into three grammatical classes: subordination ('taxi driver'), attributive/appositive ('high school') and coordinating ('poet-doctor'). These classes are based on how the relations between the words in the compound are formed syntactically. Another level of classification is semantic in nature and deals with how the semantic relations between the words are formed (this level is absent in coordinating compounds). The third level divides the compounds into endo- and exocentric. Finally Scalise and Bisetto points out that this classification leaves room for further subcategories where, for example the lexical categories of the words play a role.
Chapter 4: Early generative approaches, by Pius ten Hacken, pp. 54-77. Ten Hacken starts by reviewing works from the 60's. The general idea at that time was that compounds were transformationally related to sentences. According to ten Hacken, the specific research questions were 1. how are compounds related to underlying sentences?, and 2. how can compound formation and [semantic] classification be linked? With the introduction of lexicalism (Chomsky 1970) the general consensus seemed to be that the answers to the two questions were 1. they are not, and 2. it can't. Ten Hacken concludes the chapter with the observation that lexicalism transferred the formation of compounds from syntax to the lexical component and a number of new research questions emerged (these are not specified by ten Hacken).
Chapter 5: A lexical semantic approach to compounding, by Rochelle Lieber, pp. 78-104. Lieber's chapter is an extension of a previous analysis of exocentric compounds. According to Lieber, the semantic representation of morphemes can be divided into a semantic/grammatical skeleton and a semantic/pragmatic body. She assumes that the skeleton is made up of seven semantic features which may be +,-, or absent in any particular morpheme. These seven features are specific for English and are chosen from a universal set up of features (p. 83). The body consists of perceptual, cultural and encyclopedic aspects of meaning and has (at least) two layers. One layer consists of the universal features that are not part of the skeleton, thus supposedly stable from speaker to speaker. The other layer is purely encyclopedic and the content of this layer may vary from speaker to speaker. Lieber also proposes a list of semantic features, based on previous work on languages other than English. It is not necessary that these features be binary. Lieber goes on to show how the matching of the skeletons and bodies of the morphemes in various types of compounds gives rise to various interpretations. The main conclusions from the analysis are that exocentricity is not a unitary phenomenon (see also ch. 3) and that endocentric attributive compounds (N+N such as 'file cabinet') receive a kind of semantic default interpretation (which seems to be constrained only by context), unless the semantic features of the skeleton and body of the morphemes constrain the possible interpretations.
Chapter 6: Compounding in the parallel architecture and conceptual semantics, by Ray Jackendoff, pp. 105-128. In the first section Jackendoff outlines his 'parallel architecture' (Jackendoff 1997). In the second section, he argues that even though compounds can be freely produced and interpreted, they seem to obey other rules than syntactic ones. He then claims that compounds are relics of a so called proto-language (Bickerton 1990) and discusses how N+N compounds in English may get their interpretation. Jackendoff identifies three components that contribute to their meaning: profiling ('picking out a character in an event and designating this character as the one being referred to' (p. 118)), action modality (the difference between stage and individual level predicates, for example), and finally cocomposition (the possibility to fill in unexpressed, coerced functions expressed in nouns, for example 'finish the book' = 'finish DOING something with the book' (p. 121)). In the next section he gives examples of how these components interact in the interpretation of compounds. One of his conclusions is that the failure of previous studies to list all possible relations between the components of a compound is because some components contain embedded coercions. This embedding, which in theory is indefinite, makes the number of relations indefinite, but still systematic, since the number of coercion functions is limited.
Chapter 7: Compounding in distributed morphology, by Heidi Harley, pp. 129-144. In the first section Harley points out that compounding has not been a major concern in distributed morphology (DM). This is indeed so, but there is at least one work that treats compounds in DM, namely Josefsson 1998, which Harley does not mention. The second section is a short introduction to the basic principles of DM. Harley's conclusion is that there is basically no distinction between inflectional and derivational affixes, they are all just bundles of different kinds of features. This makes them different from roots which refer to encyclopedic content. The general idea in DM is that word formation is syntactic. As a consequence, roots are acategorial and have to be concatenated ('merged' in current syntactic terms) with category forming heads in the syntax. In her analysis Harley argues that internal arguments are concatenated to the root before it gets its categorial status. However, this way of forming compounds gives rise to the familiar 'bracketing paradox', i.e. the word 'truck driver' is formed as [[truck drive] er] rather than [truck [driv er]]. In the following sections the same analysis is given for modificational compounds ('quick growing') and root compounds. The fact that we don't get incorporation of roots into verbs (such as 'to truckdrive') is according to Harley because of parametric variation on what v-heads in different languages allow as incorporated material. In her analysis of phrasal compounds ('the bikini-girls-in-trouble genre'), she assumes that the phrase 'bikini-girls-in-trouble', XP, is merged with a n-head. XP is simply treated as a root. This nP can then incorporate in the head of the compound ('genre'), just as in other types of compounds.
Chapter 8: Why are compounds part of human language? A view from asymmetry theory, by Anna Maria di Sciullo, pp. 145-177. According to di Sciullo all types of compounds share the property of relating their component parts in an asymmetric way. She draws parallels to syntax (see Kayne 1994) and morphology (derivation and inflection). The aim of the chapter is to provide an account of compounds in asymmetry theory (di Sciullo 2003). She claims that there is a functional projection, F, that relates the parts of the compound. The head of the compound is the complement of F and the modifying element in the compound is in the specifier position of F, hence the asymmetric relation between the two components. The functional heads have one of the following meanings: 'and', 'or', 'with', 'to', and 'in'. It may sometimes be pronounced, as in 'hit-and-run'. Di Sciullo shows how compounds are derived in a special morphological component of the grammar which works in parallel with the syntax. The machinery is the same as in minimalist syntax; phases, interpretable and uninterpretable features and agree.
Chapter 9. Compounding and lexicalism, by Heinz Giegerich, pp. 178-200. Giegerich starts by giving a short outline of the history of lexicalism and its assumptions about the structure of the lexicon and the divide between syntax and morphology. By carefully analyzing a wide range of compounds Giegerich shows how two core principles of lexicalism, 'lexical integrity' (Lapointe 1980) and the 'no phrase constraint' (Botha 1984) cannot be maintained, at least not unmodified. Finally Giegerich discusses the lexical strata of English. He claims that there are (at least) 2 strata and he also makes the somewhat surprising claim that exocentric bahuvrihi compounds are not productive in the language (ch. 18 expresses another opinion).
Chapter 10. Compounding and construction morphology, by Geert Booij, pp. 201-216. The aim of the chapter is to apply the ideas of construction grammar (Goldberg 2006) to morphological processes in general and compounding in particular. Booij argues that compounds are best accounted for by means of templates of the following type [X Yi]y = Yi with relation R to X, R being specific for each individual pair X and Y (p. 203). This template is elaborated on and expanded. Booij claims that there is a morphological component in addition to syntax and phonology. In the following sections Booij shows how (mainly Dutch) compounds can be analyzed. Unfortunately the implementation is very sketchy and it is difficult to see how the templates used are more than descriptive notations.
Chapter 11. Compounding from an onomasiological perspective, by Joachim Grzega, pp. 217-232. Grzega gives very short outlines of a few onomasiological approaches to word formation. It seems that compounding has not been a major concern for any of them.
Chapter 12. Compounding in cognitive linguistics, by Liesbet Heyvaert, pp. 233-254. The first section is a run-through of some of the key concepts used in cognitive linguistics. In the following sections Heyvaert presents various approaches to compounding in the framework. She concludes with suggestions for future research.
Chapter 13. Psycholinguistic perspectives, by Christina Gagné, pp. 255-271. Gagné starts by identifying some theoretical issues in psycholinguistic research. There are, according to her, three major questions: how complex words are represented, when (if at all) decomposition occurs, and if morphology is explicitly represented. There seems to be disagreement concerning all three. Next she presents some factors that influence the processing of compound words. Data suggest that representations of both the compound and its component parts become available during processing. The conclusion Gagné draws is that the mental lexicon is highly structured with multiple levels of representation. Compounds appear to be stored so that their representations share aspects of form and meaning with other words (simplex and complex). They also take advantage of the interconnected network of relations among concepts, though the exact nature of these systems are still being explored.
Chapter 14. Meaning predictability of novel context-free compounds, by Pavol Štekauer pp. 272-297. Štekauer outlines the difference between meaning prediction in words and in compounds. He goes through some previous approaches to meaning prediction before he presents his own approach. Compounds are classified into four onomasiological types: 1: compounds that are mostly syntactic ('piano player', 'dishwasher'). 2: compounds where the actional semantic constituent is available and the object acted on is also available ('plaything'). 3: compounds where there is no unambiguous relation between the components ('baby book' a small book, a book for or about babies etc). 4: compounds of the object incorporated kind ('baby-sit', 'stage-manage').
Chapter 15. Children's acquisition of compound constructions, by Ruth Berman, pp. 298-322. After discussing the problems compounds confront children with, Berman presents data from various cross-linguistic studies. The conclusion she arrives at is that there is considerable variation in use among children and that children's use of compounds is governed by the general frequency and the stylistic level of compounds in the target language. The chapter also includes a case study of the developmental route of Hebrew. It concludes with a discussion of areas where further research is needed.
Chapter 16. Diachronic perspectives, by Dieter Kastovsky, pp. 323-340. Kastovsky identifies three types of morphological systems: word based (the basic morphological entity can function as a word in the language), stem-based (the stem is not a word without additional morphology), and finally, root-based morphology (the root may require more morphological to become a stem, as Semitic 'k-t-b' 'write'). A language is rarely of one type only, but more or less of a certain type. Kastovsky's conclusion about the historical development of compounds in Indo-European is that it goes back to the lexicalization of syntactic phrases that were used attributively and that the adjectival bahuvrihis played the most important role.
Chapter 17. Typology of compounds, by Laurie Bauer, pp. 343-356. Bauer discusses the problems of defining compounds, a problem that for obvious reasons spills over to their typology. Then he discusses various parameters where we find typological differences. The first one is universality. There are claims that compounding is not a language universal; however, the definitional problem makes such claims very hard to evaluate. Some languages seem to allow object incorporation but not compounds (Kwakwala, Štekauer et al. 2007), but according to some definitions incorporation is a sub-category of compounding. The overall typological picture of compounding is that there are hardly any generalizations that apply.
The rest of the chapters (18-34) are typological studies of individual languages. The organization is the following: Indo-European: Germanic (English by Rochelle Lieber, Dutch by Jan Don, German by Martin Neef, Danish by Laurie Bauer), Romance (French by Bernard Fradin, Spanish by Laura Malena Kornfeld), Hellenic (Modern Greek by Angela Ralli), Slavonic (Polish by Bogdan Szymanek); Sino-Tibetan: Mandarin Chinese by Antonella Ceccagno and Bianca Basciano; Afro-Asiatic: Semitic, Hebrew by Hagit Borer; Isolate: Japanese by Taro Kageyama; Uralic: Finno-Ugric, Hungarian by Ferenc Kiefer; Athapascan: Slave by Keren Rice; Iroquoian: Mohawk by Marianne Mithun; Arawakan: Maipure-Yavitero by Raoul Zamponi; Araucanian: Mapudungun by Mark C. Baker and Carlos A. Fasola; Pama-Nyungan: Warlpiri by Jane Simpson.
This is a very good book. It gives a thorough overview of the theoretical approaches to compounding and almost half of it contains typological studies of individual languages. An extra plus is the diversity of theoretical frameworks that are represented. A wide range of approaches are covered, and no linguist can be expected to be familiar with all of them, but the chapters are easy to follow. They start by introducing the concepts that will be used and from that take the reader through state of the art analyses of compounding in their respective framework.
Even though the bulk of the languages presented in the second part are Indo-European and particularly Germanic and Romance (in fact, all other language families are represented by one single language), the number of non Indo-European languages is greater than in most handbooks in linguistics.
In addition, the book is carefully edited and the chapters contain appropriate cross-references (there are some apparent ones missing, for example some authors' claims concerning universality and language specific productivity are disputed in the typological section). The only thing that stands out is the amount of time spent on the definition of compounding. The whole of the first chapter is basically a discussion of this and it feels that some of the authors could have skipped a lot of that discussion in their own chapters, especially since they all come to the same conclusion: it is very difficult to give a good definition of compounding. But on the other hand, some authors seem completely unaware of the wealth of ways to form and interpret compounds (and often these authors only deal with n-n compounds, giving further support to Scalise and Bisetto's claim that other kinds of compounds are often overlooked (ch. 3)). Their analyses seem slightly too shallow and general (for example ch. 2, 7 and 10) and pay very little attention to the problem of accounting for both syntactic and morphological properties. In my opinion this is a book that serves its purposes (a handbook for the initiated and a starting point for beginners) in an excellent way. Thanks to the large index it is easy to look up specific theoretical issues as well as specific empirical questions in the typological section. I recommend it not only to linguists interested in compounding, but also to anyone interested in the intricate interplay between morphology, syntax and semantics.
Bickerton, D. 1990. Language and Species. Chicago: University of Chicago Press. Botha, P. 1984. Morphological Mechanisms. Oxford: Pergamon Press. Chomsky, N. 1970. Remarks on nominalizations. in Jacobs and Rosenbaum (eds.) Readings in English Transformational Grammar. Waltham, MA: Ginn, 184-221. di Sciullo, A. 2003. The asymmetry of morphology. in Boucher (ed.) Many Morphologies. Somerville, MA: Cascadilla Press, 1-33. Goldberg, A. 2006. Constructions at Work. Oxford: Oxford University Press Jackendoff, R. 1997. The Architecture of the Language Faculty. Cambridge, MA: MIT Press. Josefsson, G. 1998. Minimal Words in a Minimal Syntax: Word formation in Swedish. Amsterdam: John Benjamins. Kayne, R. 1994. The Antisymmetry of Syntax. Cambridge, MA: MIT Press. Lapointe, S. 1980. The Theory of Grammatical Agreement. PhD Dissertation, University of Massachusetts, Amherst. Štekauer, P, S. Valera and L. Körvélyessy. 2007. Universals, tendencies and typology in word-formation. Cross-linguistic research. Ms.
ABOUT THE REVIEWER
ABOUT THE REVIEWER:
Fredrik Heinat currently holds a research position at Stockholm University.
He is involved in a project on complex predicates, light verbs and argument
structure. The focus of his research is on the relation between morphology,
syntax and semantics.