Along with a number of cognitive linguists (Schmid 2011, Hamawand 2011, Ungerer 2007) we presume meaning to have decisive relevance in derivational morphology. As Hamawand (2011:15) points out: “Prefixes ... have meaning of their own, which contribute to the semantic import of the host roots. [] In word-formation, a prefix is the most important part because it lends its character to the whole derivative. This is so because it adds to the derivative a new shade of meaning.” The very topic of Hamawand’s essay concentrating on prefixal rivalry points at the essence of the outlined problem of how the semantic content of prefixes is treated. Prefixal rivalry refers to the possibility of an alternation between two or more prefixes playing a part in creating derivatives from the same root. This is especially relevant in cases of prefixes of degree e.g. ultra-confident and super-confident (Hamawand 2011:14). Though both forms are derived from the adjective confident, they show difference in use.2 As Hamawand claims, though two rival prefixes might describe a conceptual content represented by the root, both do so in their characteristic way. Within the theoretical tenets of the cognitive linguistic paradigm, it is meaning which counts as a decisive element in the choice between rival prefixes. Instead of homonymy, prefixes of degree have multiple meanings.
2 In The captain is ultra-confident the adjective ultra-confident means “The captain is confident to an undue degree”. In The captain is super-confident, the adjective super-confident means “The captain is confident to an intense degree”. Hamawand (2011:14)
84
Schmid (2011:147) also states typical prefixations to be combinations of free lexical morphemes serving as bases, with preceding bound lexical morphemes (the prefixes themselves) in a modifying function. Prefixes therefore form paradigms, that is, they operate on definable types of bases. As a rule, in such typical prefixations the word class of the head determines the word class of the whole derived complex lexeme as well.
From a lexicalizational point of view, typical prefixations are weakly lexicalized, as both semantic and formal lexicalizations lead quickly to the loss of transparency. Not all prefixations, however, follow this typical prefixational pattern. Schmid (2011:148) distinguishes between two classes of the so-called untypical prefixations. Synthetic prefixations are termed such complex lexemes which are not immediately analyzable, as their bases are not existing English lexemes3. Besides the synthetic prefixations Schmid also distinguishes the subcategory of pseudo-prefixations, post-Freudian serving as an example where we do not face any segmentational problems, the morphological analysis though is to an extent opposed to the semantic analysis (the prefix only refers to a part of the base: Freud, not Freudian). The problem of pseudo-prefixations and prefixoids is amply dealt with e.g. the German literature on prefixes. Although it is indeed possible to create a list of features supporting the evidence for such a subcategory, as several authors emphasize (Brdar-Brdar-Szabó, 2000) the introduction of further subcategories does not solve the inherent classificational problems. The classificational problems concerning the category ‘prefix’ are evident. The cognitive linguistic view is particularly susceptible to notions of fuzzy boundaries or the lack of clear dividing lines. The root of the problem ─ as it is poignantly exemplified in cases of creative prefixations─ lies in judgments concerning morphological autonomy as it is also shown by the blurred dividing line between combining forms and prefixes. The question arises then, where to place the specific cases of prefixations, termed creative prefixations within the categorisational schemata? The term ‘creative’ concerning complex lexemes was first introduced by Benczes (2006) focusing on noun-noun compounds, the meanings of which can be traced back to metonymical and metaphorical processes. Creative prefixations do represent a special category, especially though from a semantic point of view: notions of prefixation based on simple rules of compositionality unfortunately fail in capturing the meaning of such prefixations. Examples like unhave, de-bachelorise or defriend cannot be understood as a sum of their respective constituents, though they are immediately analyzable, as opposed to synthetic prefixation. In fact, we must presume the emergence of a whole new concept as a result of such derivational processes. Creative prefixations therefore are a culmination of notions questioning clear-cut semantic roles assumed by the constituent members of derivational processes, placing much more importance on the semantic content of the prefixes themselves4. This semantic content is the main reason why such cases can be dealt with within the framework of Cognitive Linguistics. Cognitive linguists dealing with prefixation (Ungerer 2007, Schmid 2011, Tolcsvai Nagy 2010) ask the question, what cognitive function is served by the evidence of this word-formation pattern. From a broader perspective, it is worthwhile to mention the cognitive aspects of word-formation itself. As Onysko and Michel (2010) emphasize, within Cognitive
3Undoubted as adjective and the derived undoubtedly are not formed in the bases of doubted or *to undoubt. (Schmid 2011:148)
4 Ungerer (2007: 651) based on Langacker (1987:12) in fact states the very task of Cognitive Linguistics as the semanticization of word-formation analysis.
85
Linguistics, the cognitive aspect of word as representing a conceptual unit emerges as a decisive criterion. This is the most striking difference as opposed to both the structuralist and the generativist tradition dealing with morphology/word-formation: both approaches work with the unifying basic principle of the presumption that word-formation is basically rule-governed. Bauer (1983:293) also states that word-word-formation is rule –governed, admitting though that these rules are complex and far from obvious.
Within the same vein, Plag (2003:47-51) also emphasizes the inherent rule-governance within word-formation, yet implies a cognitive perspective on processes of word formation in word storage and productivity.
The cognitive linguistic view sees the profiling of a contrast or the basic concept of
‘different from X’ as the essence of prefixation, based on our general cognitive ability of making comparisons. Schmid (2011:160) however, as a further specification, deems the function of prefixation to be the encoding of contrasts first and foremost by lexical, rather than grammatical, means. This emphasis on the lexical signifies an even further step from a purely grammatical, hence compositional viewpoint5, in the direction of postulating specific semantic content for prefixes and prefixations, although several authors make note of semantic inconsistencies concerning prefixation (Bauer 1983, 2003, Lehrer 1999).
Moreover, it has also been observed that contrast, in case of e.g. negative prefixation cannot be sufficiently defined by the ‘not X’ definition of Matthews (1974)6, as unhappy for example is not a clear contrast of happy, implying a scalar approach, where unhappy is somewhere on the line between happy and a number of lexicalized possible contrasts. In the same lines, the prefix re- poses similar problems: it is typically used with verbs expressing accomplishment that is the verb concerned implies that the action has once already been finished. Re-, in the prototypical meaning of repeated action implies a contradiction, as Schmid (2011:160) pointedly states, there does not seem to be much sense in repeating something which has already been completed. If we take the example of rewrite, the interpretation of producing a better, edited or even re-structured version of an originally completed piece of writing seems much more plausible than making actually a copy of it.
Contrastivity, as a unifying notion concerning prefixation paves the way for several differing cognitive approaches, as summarized in Ungerer (2007:658-660). One of these is the so-called profiling contrast, expressing a special type of interpretation of the basic Figure-Ground distinction (also and perhaps even more widely referred to as trajector-landmark relationship in the literature). The unifying idea of contrast can also be backed by mapping the productive gaps in prefixation. There are verbs which typically do not permit contrast: *unlive or *unsleep, and also most concrete nouns. This is where the first important creative aspect arises: as creative examples we do have to unhave and to
5 Langacker (1987:452,461) proposes a scaffolding metaphor to implicate that in case of complex structures the respective components of word-formation structures cannot and should not be understood as building blocks of the complex structures.
6 Concerning the basic schematic structure of Matthews, it needs to be acknowledged though, that in fact there are cases, where this schema works: in case of binary contrasts like even-uneven (numbers) the schema is perfectly justified.
86
defriend.7 In both these cases we can postulate a certain violation or rules, where defriend, besides exemplifying a category-changing prefix, also displays analogical processing on the basis of befriend. The reason why we cannot form prefixations with stative or nonconclusive durative verbs respectively is that “...they do not have any natural contrast. For the same reason nouns that describe concrete objects are not suitable as bases for prefixations, since they do not have any obvious opposites.” (Schmid 2011:161)8 It is important to note that contrastivity should not necessarily be dealt with as a unified cognitive principle, but rather as a complex category, which is rooted in several image schemas. (Ungerer 2007 quoting Mettinger 1994, 1996). Adjectives prefixed with un- may involve two image schemas: the SCALE schema (e.g. unimportant) and the CONTAINER
schema for adjectives which are contradictorily negated. Still within this reasoning, the un- predication would be constituted by the relation between the “extraposed” trajector (the prefix itself) and the landmark.
The cognitive linguistic view of prefixation as conceptual reprofiling reflects upon the extra semantic information that cannot be captured by building-block compositionality. It emphasizes the condensed nature of prefixations as well as a potential syntactic simplification. The most important aspect is, however, that a “state of affairs is moulded into a concept; an idea is entrenched and can be stored in the mental lexicon.” (Schmid 2011:162)