In this contribution, I offer a summary of my 2014 Ph.D. dissertation from the University of Patras on headedness in word formation and lexical semantics.
A basic notion in morphological theory and especially in frameworks which attribute internal hierarchical structure to words is the notion head. Since the seminal works of Lieber (1980), Williams (1981), Selkirk (1982), and Di Sciullo & Williams (1987), the syntactic notion of head has been extended from syntax to morphology. A corollary of this is that words, like syntactic phrases, are considered to have heads and thus the identification of head in all morphological configurations, derived words, compounds, and inflected forms, has been central to linguistic morphology. This extension, however, has not been without problems and several authors have expressed their scepticism regarding the overall use of this notion (see for instance the work of Bauer 1990).
In recent years there has been an upsurge of interest in the study of head and related issues, such as the presence and absence of head, and several approaches have been developed (Bauer 2008, Scalise & Fábregas 2010, Kageyama 2010, Ralli 2013, Ralli & Andreou 2012).
The application of the notion of head to morphological analysis, however, is highly problematic since ‘head’ has been used in very different and often confusing ways. As far as semantics is concerned, the head is usually identified by the hyponymy test which renders the hyperonym head of the word. ‘Head’ in lexical semantics, however, has never been studied in any detail despite the fact that the hyponymy test is not particularly useful for the identification of head in a number of morphological configurations. In addition, our understanding of the head-dependent linearization and the factors–system-internal or system-external–which govern this linearization, is rather limited.
2. The Study
2.1. Aims of the Thesis
The thesis aims to sort out some of the confusions associated with the notion of head, focusing on headedness in Word Formation and Lexical Semantics. In particular, the purpose of my thesis is to inquire into the notion of head with a focus on the following three issues:
presence and absence of a head in morphological configurations.
Given that the thesis aims to provide a comprehensive study of headedness, data comes from a number of languages and from different evolutionary stages. For example, in order to delimit ‘head’, one must take into consideration a number of morphological phenomena, such as compounding, English prefixation, and Greek evaluative suffixation. In addition, the question of position of the head necessarily involves the presentation of data from previous evolutionary stages of the Greek language, such as Ancient Greek.
The thesis also aims to contribute to the study of Modern Greek dialects. In particular, I focus on data from two peripheral dialects, namely Italiot and Cypriot Greek. The analysis of morphological phenomena in these dialects can greatly inform the general discussion of headedness. In particular, the analysis of Italiot compounds allows one to comment on the head-dependent linearization, since in this dialect one finds left-headed compounds despite the fact that Greek compounds are generally right-headed. In a similar vein, Cypriot is particularly rich in compounding patterns and exhibits some of the rarest attested types of compounds, that is, verbal bahuvrihis. The analysis of compounding in Italiot and Cypriot is largely based on the rigorous descriptive studies of Andreou & Koliopoulou (2012) and Andreou (2013).
2.2. Structure of the Thesis and Main Proposals
The thesis comprises four parts with two chapters each and a final chapter which summarizes the discussion. The nine chapters are followed by two appendices. Appendix A contains Greek left-headed [N N] N compounds and Appendix B presents an overview of the history of the distinction between endocentric and exocentric compounds and shows that credit for the introduction of the term ‘exocentricity’ into linguistic theory is not due to Bloomfield (1933).
2.2.1. Introducing Head
Part I serves as an introduction to the aims and structures of the thesis (chapter 1) and to the issues that concern us in the thesis. In chapter 2, I broadly set out the background and motivation for the investigation. An overview of the way head has been used in morphology and syntax reveals that there are widely divergent views with respect to the definition and properties of head. The presentation of the various definitions of ‘head’ (see among others Zwicky 1985, Plag 2003, Booij 2007), shows that this lack of agreement partly stems from the fact that ‘head’ is usually defined in such a broad way that the application of this notion to morphological analysis is rendered highly problematic. For most scholars, for example, the notion ‘head’ captures the intuition that in a configuration comprising two elements, one of the elements is considered more important than the other, but this is of course a broad description rather than a definition of ‘head’. In addition, no consensus has been reached with respect to the properties of a head, and ‘head’ has been used as a notion which unifies a number of other notions, including amongst others the subcategorizand and the morphosyntactic locus.
2.2.2. On the Definition of ‘Head’
In Part II, I address the question of delimitation and definition of ‘head’. In chapter 3, I comment on the delimitation of ‘head’ in morphology with a focus on derivation and compounding. In order to do so, I first present the criteria used in the relevant literature with respect to the identification of the head and I then apply these criteria to derivation and compounding, respectively. The considerably vexed picture which emerges from the application of the various headship criteria to word formation reveals that most of the assumed head-like notions, such as the subcategorizand, might very well not be relevant to the head-nonhead asymmetry. The discussion of the results of the various headship criteria argues for a strictly categorial definition of ‘head’, in that the head for the purposes of morphology should be identified with the category determinant.
The purpose of chapter 4 is to investigate the way the head-nonhead asymmetry should be accounted for in Lexical Semantics. To this aim, I adopt the framework of Lexical Semantics as developed by Lieber (2004). In this chapter, I argue that the head for the purposes of Lexical Semantics should be a matter of the skeletal features of a morpheme and, more specifically, the head should be identified with the ontological class determinant. In addition, the application of the notion ‘head’ to the creation of semantically complex configurations reveals that the mechanism of subordination of functions should be split into subordination of functions with, (1a), and without indexation of arguments, (1b). In particular, non-argument-taking inflection, prefixation, and evaluative morphology should be accounted for by the mechanism in (1b):
This has implications for the way we classify morphemes in lexical semantics since any classification should take into consideration the following two main criteria: (a) the semantic features which are relevant to each morpheme and (b) the internal organization of the lexical-semantic representation of a morpheme. In this respect, the distinction between prefixes and suffixes which is based on the linear order of morphemes cannot be used as a criterion in lexical semantics.
2.2.3. On the Position of the Head
Although ‘head’ was introduced into morphology in the early 80’s, there is very little in the literature to date that deals with the issue of the head-nonhead order inside words. Part III aims to present system-external (chapter 5) and system-internal (chapter 6) factors which govern constituent order in morphology with a focus on Greek compounding. More specifically, although Greek compounds are generally right-headed, in chapter 5, I present and analyze a number of left-headed compounds from various evolutionary stages of the Greek language and its dialects (with focus on the dialects spoken in Southern Italy). Consider the following left-headed compounds from Italiot Greek:
First, I comment on whether the presence of these formations could be attributed to the contact of Greek with Italian, which exhibits left-headed compounds, and then I present a number of factors, such as typological dissimilarity, that militate against the language interference hypothesis.
In chapter 6, I delve more deeply into those system-internal factors which could affect the head-dependent linearization. In particular, I show that left-headedness manifests itself in previous evolutionary stages of the Greek language and other Modern Greek dialects as well. In addition, I show that left-headedness is not limited to compounds of the sporomáratho type, as it manifests itself in verb-first compounds of the type arche-laos (“lead-people”) ‘leading the people’, Homeric prepositional compounds such as en-oikios (“in-house”) ‘(one who is) in the house’, and compounds composed of an adjective with governing properties as in iso-theos (“equal-god”) ‘equal to the gods’. The analysis of left-headed patterns reveals that the order of constituents in compounds may not be autonomous from syntax, given that the head-nonhead linearization inside compounds cannot change without previous change in the head-nonhead order in syntactic constructions (Harris & Campbell 1995, Gaeta 2008). This conclusion allows one to comment on the validity of the Lexical Integrity Hypothesis and the idea that the order of elements inside compound words obeys morphological settings only (Bresnan & Mchombo 1995). Finally, the exceptionality of the phenomenon of reordering and variation in compounding is highlighted by the fact that it has no parallel in derivation and inflection.
2.2.4. On the Presence and Absence of ‘Head’
The presence or absence of ‘head’ cuts across the endocentricity and exocentricity distinction but in order to address the question: The presence or absence of which head? Part IV aims to offer a detailed investigation of a phenomenon which, in my opinion, partly reflects the problematic way scholars have treated the notion ‘head’.
Chapter 7 serves as an introduction to the distinction between endocentric and exocentric compounds. In this chapter, I critically evaluate literature on this issue and present data that militates against recent proposals that exocentricity and the notion head can be split into morphological, semantic, and categorial. Based on the distinction between nominal and adjectival bahuvrihis, which is often not taken into consideration by scholars, I argue that the former should be analyzed via metonymy and that the latter can be better understood if we examine the relation between compounding and derivation.
In Chapter 8, I focus on the relation between two word-formation processes: compounding and derivation. In particular, I argue that the head in exocentric compounds is a derivational suffix (overt or zero) and present evidence in favour of zero-derivation. To adduce an example, Greek verbal bahuvrihis, such as the Cypriot varikart-izo ‘to have a hard heart, to be sad’, which are the rarest type of bahuvrihi compounds, are built on the basis of the pattern [[stem stem]-Dsuf]. In addition, I tackle the issue of the inconsistency between the structure suggested by the morphophonological properties and the structure suggested by the meaning of exocentric compounds (bracketing paradoxes). Finally, I comment on the distinction between de-compounds and exocentric compounds and argue that exocentricity is compounding for the purposes of derivation.
Chapter 9 summarizes the main proposals and looks ahead to possible advances.
In a nutshell, the main proposals with respect to headedness in morphology and lexical semantics are as follows:
Most of the assumed head-like notions, such as the subcategorizand, the morphosyntactic locus, and the obligatory constituent, might very well not be relevant to the head-nonhead asymmetry.
The head for the purposes of morphology should be identified with the category determinant and with the ontological class determinant as far as lexical semantics is concerned.
The enquiry into the postulation of functions and arguments reveals that non-argument taking inflection and derivation should be accounted for by the mechanism of subordination of functions without indexation of arguments. This has implications for the way we classify morphemes in lexical semantics.
With respect to the position of head, the phenomenon of left-headed [N N] N compounds in Italiot Greek should not be studied independently of the same phenomenon evident in previous evolutionary stages and in other dialects.
In addition, morphology does not combine and manipulate asymmetric relations only (see Di Sciullo 2005), given that morphemic order is amenable to change and variation may arise inside the morphological system of a language.
From a diachronic point of view, the order of constituents in compounds may not be autonomous from syntax, since the head-nonhead linearization inside compounds cannot change without previous change in the head-nonhead order in syntactic constructions.
With respect to the absence of a head, the analysis of the compound patterns which are considered to be exocentric reveals that exocentric compounds are bracketing paradoxes which involve compounding and derivation in this particular order. In this respect, I part company with previous analyses which identify exocentricity based on the hyponymy test and which collapse exocentricity and semantic non-compositionality.
To conclude, my work on the notion ‘head’ in word formation and lexical semantics raises a number of secondary questions, some of which I address in my thesis. I hope that future work will delve more deeply into the study of these questions.
Andreou, Marios. 2013. Compounding in the Greek dialects of Southern Italy. In Mark Janse, Brian Joseph, Angela Ralli & Metin Bagriacik (eds.). Proceedings of the 5th International Conference on Modern Greek Dialects & Linguistic Theory, 1–18. Patras.
Andreou, Marios & Maria Koliopoulou. 2012. I sinthesi stin Kipriaki ke Kalimniki: Mia proti sigkritiki proseggisi. [Compounding in Cypriot and Kalymniot: a first comparative analysis]. Neoelliniki Dialektologia 6: 7–29 (Athens: Academy of Athens).
Bauer, Laurie. 1990. Be-heading the word. Journal of Linguistics 26: 1–31.
Bauer, Laurie. 2008. Exocentric compounds. Morphology 18: 51–74.
Bloomfield, Leonard. 1933. Language. New York: Holt, Rinchart and Winston.
Booij, Geert. 2007. The grammar of words: An introduction to linguistic morphology (2nd ed.). Oxford/New York: Oxford University Press.
Bresnan, Joan & Sam Mchombo. 1995. The lexical integrity principle: Evidence from Bantu. Natural Language and Linguistic Theory 13: 181–254.
Di Sciullo, Anna Maria. 2005. Asymmetry in morphology. Cambridge, MA: MIT press.
Di Sciullo, Anna Maria & Edwin Williams. 1987. On the definition of word. Cambridge, MA: MIT Press.
Gaeta, Livio. 2008. Constituent order in compounds and syntax: typology and diachrony. Morphology 18(2): 117–141.
Harris, Alice C. & Lyle Campbell. 1995. Historical syntax in cross-linguistic perspective. Cambridge: Cambridge University Press.
Kageyama, Taro. 2010. Variation between endocentric and exocentric word structures. Lingua 120(10): 2405–2423.
Lieber, Rochelle. 1980. On the organization of the lexicon. Ph.D. Diss., Massachusetts Institute of Technology, Cambridge, MA. [Published by Indiana University Linguistics Club, 1981 and Garland Press, 1990.].
Lieber, Rochelle. 2004. Morphology and lexical semantics. Cambridge: Cambridge University Press.
Plag, Ingo. 2003. Word-formation in English. Cambridge: Cambridge University Press.
Ralli, Angela. 2013. Compounding in Modern Greek. Dordrecht: Springer.
Ralli, Angela & Marios Andreou. 2012. Revisiting exocentricity in compounding: evidence from Greek and Cypriot. In Ferenc Kiefer, Mária Ladányi & Péter Siptár (eds.). Current issues in morphological theory: (ir)regularity, analogy and frequency, 65–81. Amsterdam/Philadelphia: John Benjamins.
Scalise, Sergio & Antonio Fábregas. 2010. The head in compounding. In Sergio Scalise & Irene Vogel (eds.). Cross-disciplinary issues in compounding, 109–126. Amsterdam/Philadelphia: John Benjamins.
Selkirk, Elisabeth. 1982. The syntax of words. Cambridge, MA: MIT Press.
Williams, Edwin. 1981. On the notions of ‘lexically related’ and ‘head of the word’. Linguistic Inquiry 12: 245–274.
Zwicky, Arnold M. 1985. Heads. Journal of Linguistics 21(1): 1–29.
I would like to take this opportunity to thank my Ph.D. committee members, Prof. Angela Ralli, Prof. Rochelle Lieber, and Prof. Brian D. Joseph, for their encouragement and guidance. This research has been funded by the ‘Constantin Carathéodory’ program (D.159) of the University of Patras.