Separation Hypothesis

by Eva Forster (Author)

Term Paper 2005 17 Pages

English Language and Literature Studies - Linguistics



1. Introduction

2. Lexeme-based theories vs. morpheme-based theories
2.1. Morpheme-based theories: Lexical Morphology
2.2. Morpheme-based theories: Lexeme/Morpheme base Morphology (LMBM)

3. The Separation Hypothesis (SH)
3.1. The starting point of SH Polyfunctional affixes Zero-morphology and empty morphs
3.2. SH – an outline Lexical Derivation

4. Arguments in favour of SH

6. Problems with SH

7. Conclusion

8. Bibliography

1. Introduction

Morphology, which can be roughly defined as "the study of words and their structure" (Bauer 2003: 3), is one part of linguistic theory, which had to fight for its justification the hardest in the middle of the 20th century, even though it is "one of the oldest concerns of linguistics" (Aronoff 1983: 355). Yet, after its heyday in the nineteenth century, it slowly was marginalized and lost its supreme status, the major reason being the rise of generative linguistics[1], which mainly focused on syntax and phonology. This change of attitude to morphology paradoxically can be explained by its central position it has within linguistic theory. Since it is concerned with words, which

- obviously have phonological properties,
- can be put together to form sentences,
- in many cases have a special syntactic function and
- are composed of parts, which again can have a meaning of their own,

morphology is situated at the interface of phonology, syntax and semantics. (cf. Spencer, Zwicky 2001: 1). As a consequence some linguists started to consider morphology as a part of phonology or syntax, denying it an independent status.

However, when the first enthusiasm in the new field of syntax had abated, linguists began to recognise the implications, which discoveries in the field of syntax and phonology had for morphology. This is one of the reasons why morphology from around the 1970s on regained some of its earlier importance, even though it "has not generally managed to establish itself alongside [phonology and syntax]" (Carstairs-McCarthy 1992: 3). Since then a great variety of approaches to morphology have been developed (e.g. Lexicalist Morphology, Word-and-Paradigm Morphology, Parallel Morphology, Meaning-based morphological theories such as Lexeme-Morpheme-Based Morphology, Natural Morphology, Distributed Morphology and many more).

In the following I will concentrate on one special idea in theoretical linguistics and its implications for morphology – the Separation Hypothesis, developed by Robert Beard (1966-1995). Beard tried "to find a universal set of principles governing all meanings expressed morphologically, […] and claims that the semantic side of morphology obeys principles which are to a large extent independent of its formal, or morphophonological, side." (Carstairs-McCarthy 1992:173) As a consequence his approach is a rigid separation of the two aspects of morphology – form and meaning (therefore the name of his hypothesis, on which he built his morphological framework).

The aim of this paper is to show how the Separation Hypothesis works and to elaborate on the main arguments in favour of this hypothesis and the problems that might occur when dealing with morphology in this way. To highlight its unique status and underline its main ideas it will be contrasted with Lieber's lexical approach to morphology. Yet, this approach will be introduced only as far as it is needed to accentuate the characteristics of Beard's theory. As the very short glance on morphology as a subdiscipline has shown, this field of linguistic theory is, due to its status as an interface, an ideal area for divers investigations. It goes without saying that the Separation Hypothesis, the basis of Lexeme-Morpheme-Base Morphology, could be contrasted with numerous other theories different or similar to it, in order to set it apart. Nevertheless, this paper will focus only on Lieber's lexical approach to morphology in opposition to Beard's, since a) Beard himself uses her theory as a contrast in his papers and b) any other outline simply would go beyond the scope of this paper.

2. Lexeme-based theories vs. morpheme-based theories

Among the different approaches to morphology one can differentiate between at least two ways of looking at the field of activity of morphology. Aronoff distinguishes lexeme-based morphology from morpheme based morphology in particular, stating that the latter "reduce[s] language to simplex signs, each of which is an arbitrary union of sound and meaning". In contrast lexeme-based morphology "starts from two decidedly unstructuralist assumptions: that the morpheme is not the basic unit of language and that morphology and syntax are not one and the same." (qtd. In http://www.facstaff.bucknell.edu/rbeard/lexbase.html).

Morpheme-based morphology, in other words, assumes that language contains only one type of meaningful unit, the morpheme, which includes stems and affixes, all of which are signs.

Lexeme-based morphology assumes that only lexemes, derived or underived, are signs, and that affixes, reduplication, revowelling, metathesis, subtraction, stem mutation, and the like, are means of phonologically marking independent derivational operations which a lexeme might have undergone.

Lexeme-based morphology, it follows, assumes the Separation Hypothesis, that assumption that the derivation of meaning and the realization of phonological marking are distinct processes in word-formation. Advocates of morpheme-based morphology include Lieber (1992), Bresnan (1982), DiSciullo and Williams (1987), Selkirk (1982), and Scalise (1986). The leading advocates of lexeme-base morphology are Aronoff (1976, 1994), Anderson (1992), Beard (1966-1995), Halle and Marantz (1993), Zwicky (1989) and Stump (1991). (Beard: http://www.facstaff.bucknell.edu/rbeard/lexbase.html)

2.1. Morpheme-based theories: Lexical Morphology

In 1970 Chomsky lay the foundation for Lexicalism with his "Remarks on nominalization". He argued that word-formation processes (especially derivation) are different from those occurring in syntax and therefore have to happen in the lexicon. Hence, some linguists started to think of the lexicon[2] not only as of a list containing "those items which are irregular in some way" (Aronoff 1983: 358), but figured, that it must be a an active component including lists of morphemes, underived words and rules for word-formation.

According to Beard (1987) the basic theoretic assumptions of Lexical Morphology are the following:

1. Morphemes (stems and affixes) are linguistic signs.
2. Morphemes are stored in the lexicon (Lexical Affix Hypothesis (LAH)[3] ).
3. In the lexicon all of the morphological and phonological operations are carried out.

Morphological frameworks working with these kind of postulates have to face quite different aspects of morphology – form, syntax and meaning. Dealing with this hybridity leads to several problems, which will be discussed in the next chapters, when Beard's ideas about how one should handle word-formation, are presented.

As a proponent of Lexical Morphology (cf. Beard 1987: 3) and advocate of Word Syntax[4] (cf. Beard 2001: 47) Lieber generally agrees with the points enumerated above. In her lexicon one can find unstructured information about listemes, such as phonological and semantic representations[5] and the syntactic category to which they belong. (cf. Lieber 1992: 21-22) She further distinguishes bound morphemes from free morphemes, by giving them an additional entry that indicates the category items to which they may attach and which they produce. As a consequence, for Lieber, affixes and no derivation operations determine the category of newly derived words. (cf. Beard 1987: 2-4)

2.2. Morpheme-based theories: Lexeme/Morpheme base Morphology (LMBM)

As cited above, Aronoff defined lexeme-based morphological theories as approaches to morphology which consider only lexemes to be linguistic signs. Lexeme/Morpheme Base Morphology (LMBM) is a variant of these kind of theories. Thus, the crucial difference between LMBM and other morphological frameworks, especially Lexical Morphology, is the strict distinction of lexemes and morphemes.

For this reason Beard has to find another definition for the morpheme, since he does not use it in the traditional Bloomfieldian sense anymore. For Beard, morphemes have no meaning whatsoever and therefore are no signs, but just a marking signifying that a derivation has taken place. So the definition of the morpheme reads as follows:

"The smallest phonologically distinguishable bound or unbound formal modification of a lexeme which is paradigmatically ordered in a closed class, marks grammatical functions, and is not susceptible to L-derivation[6] ". (Beard 1988:44)

On the contrary, lexemes (which Beard defines explicitly as noun, verb, and adjective stems) are the elements that have meaning and therefore are signs in the Saussurian sense. A sign is, according to Saussure, a mutually implied arbitrary pairing of sound and meaning. This mutual implication naturally suggests that changes in sound provoke changes in meaning, and if a speaker wants to express another meaning he or she has to use another sound. (cf. Beard 1987: 2) Hence a lexeme is defined as "a minimal, distinguishable bound or free ordered sequence of phonological segments associated with at least one sense and extrinsic referent, which may be lexically extended by L-derivation and which belongs to an open unparadigmatic class." (Beard 1988:44)


[1] Noam Chomsky's "Syntactic Structures" was published in 1957.

[2] This so called Lexicalist Hypothesis (LH) was actually never practiced in such a radical way, rather most theories of word formation, which are based on the LH "argue for some degree of interaction between morphology and syntax." (Lieber 1992: 19)

[3] The Lexical Affix Hypothesis goes back to Bloomfield, a structuralist linguist, who introduced the term morpheme in the 1930s, defining it as the smallest meaningful unit one can find in a language. He distinguishes between 'bound' morphemes (i.e. affixes) and 'free' morphemes (i.e. lexemes), drawing from this the conclusion that "all morphological signs and the rules which operate on them are proper to the lexicon alone." (Beard 1987: 2)

[4] Word Syntax applies syntactic knowledge on the structure of words. (cf. Bußmann 2002: 756)

[5] Lieber calls these representations Lexical Conceptual Struture (LCS).

[6] Derivation, or lexical derivation, is in Beard's terminology "the semantic relationship between a simple open-class lexeme such as bake and a related complex word such as baker." (Carstairs-McCarthy 1992: 182)


ISBN (eBook)
ISBN (Book)
File size
543 KB
Catalog Number
Institution / College
University of Stuttgart – Linguistik-Anglistik
Separation Hypothesis Advanced Morphology


  • Eva Forster (Author)

    9 titles published



Title: Separation Hypothesis