10.5.4. Pathemic semes

We now need to ask to what extent this level of meaning comes into a description of the linguistic meaning of individual tetms. Is lugubriousness a semantic prop­erty of black? I believe that this problem, and similat ones in linguistic semantics, must be resolved on a linguistic basis. The deciding criterion should be, in my view, the following: whete conventionalized extensions of a certain property are tegistered in the semantic system of a given language,' this property should be part of the meaning of the term in question. In fact, we have a great many metaplmri cal and dysphoric extensions of black. Considet expressions like:

My future looks black. I am in a black mood. Black humor

in which the meaning of the adjective is not a chromatic quality, but an emotive tonality. None of these expressions could be understood if ibis tonality were not registered in the semantics of the term and did not constitute an integral pari ol competence. Once again, llie lesl for verifying the adequacy ol ibe semantic rep resentaiion is comprehension anil the practical process ol interpretation.

236

The phoric component of value can be directly inscribed in the lexical mean­ing of many terms, as in the pairs slim/skinny, or thrifty/stingy. Their semantic difference can only be explained by drawing on an axiological component based on the euphoric/dysphoric opposition. It is important to stress the inadequacy in these cases of a classic connotative model which would account for these phe­nomena in terms of additional connotations superimposed onto the basic mean­ing. Valorization is here a primary element of semantic organization, not a second­ary meaning that could be erased, and is conveyed by specific semes that we could call pathemic semes.

In other cases, lexical units include in their meaning a pathemic compo­nent whose euphoric or dysphoric axiology derives from an element that could be called enunciative perspectivization. Think of the difference between noun phrases like a violent bbw and a painful blow. An implicit point of view is in­scribed in each expression regarding the subject of the enunciation: in the first case the blow is oriented according to the perspective of the person who inflicts it, in the second to that of the person who receives it. Pathemic aspects of this kind can often be found at a lexical level, above all in qualifying adjectives; frightening, terrifying, and terrible, to give just a few examples, qualify an event or entity, but at the same time presuppose a subject who experiences the feeling of fear or terror, thus constructing an implicit relation of subject and object. If something is terri­fying or frightening, it must necessarily be so in relation to an implicit subject who feels that emotion and who is textually inscribed by that lexical choice.

io.$.$. The heterogeneity of semes and the stratification of content

The analysis carried out thus far regarding the different nature of the properties comprising the semantic universe of a language reveals an intrinsic heterogeneity of constitutive semes, which not only include different qualities, but seem also to refer to different levels of pertinence and depth. To put it in more Hjelmslevian terms, the substance of content is stratified according to a hierarchy of levels,

whose extreme levels (which are also the most important and noted ones) are the physical level on the one hand and the level of apperception and evaluation or collective appreciation on the other. (Hjelmslev [1957] 1959)

The most interesting aspect of the stratification of content is the priority of the axiological or evaluative level, which is stressed by Hjelmslev on numerous occa­sions.

Clearly, it is evaluative description that imposes itself before all else in the sub­stance of the content. It is not through a physical description of signified things that one can usefully characterize the semantic use adopted by a linguistic com­munity and belonging to the language that one wants to describe; on the con­trary, this can be obtained through the evaluations adopted by this community.

237

the collective appreciations, social opinion. (Hjelmslev [1954] 1959) It is the level of collective appreciation which constitutes the constant presup­posed (selected) by the other levels, including the physical level (which, as we know, may be absent), and which by itself enables (amongst other things) a scientifically valid account of "metaphor." (Hjelmslev [1957] 1959)

Metaphors in effect constitute an important area for verifying the priority and immediacy of the evaluative level which, as I argued earlier, is directly pertinent from a linguistic point of view, and the only one, according to Hjelmslev, which is simultaneously important on both the linguistic and the anthropological planes. The priority of value is also sustained by Greimas, albeit not on the anthro­pological and social plane that Hjelmslev focuses on. According to Greimas, the thymic may be thought of as a precondition of meaning, and its priority is onto-logical, in relation to the very possibility of meaning. It constitutes the deepest level of semanticism, affecting the fundamental, elementary structures of signifi­cation with its euphoric and dysphoric values. This position is very similar to that of Gestalt psychologists who claimed that tertiary properties have priority over all others. (According to Wertheimer, black is lugubrious prior even to being black.) At the very foundation of meaning, at its deepest level, prior perhaps to any con­vention and code, we find a pulsional intentionality made up of emotions and sensations rooted in our corporeal, perceptual, and psychic organization and in the valencies which, perhaps already inscribed in the forms of the natural world, color our world with values, affects, attraction, and repulsion.

10.6 Semantic structure and linguistic classes

As yet largely unexplored is the relation between the articulation of the semantic system and its possible differentiated distribution in the various parts of speech. Is thete a regularity between items of expressable content and a tendency for them to be codified in certain linguistic classes rathet than others? In other words, do linguistic classes have a semantic basis? And to what extent can models which ate valid, for example, for nominal semantics, be extended to other parts of speech? As far as the first of these questions is concerned, the general tendency in linguistics and cognitive semantics is certainly oriented toward the search for a semantic motivation for linguistic categories; a number of important contribu­tions have already been made to this end, especially in terms of the different distribution of content between open classes and closed classes. The important work of Talmy (1983, 1988a) has clearly shown that there are precise limits on the type of semantic information conveyed respectively by the lexicon and by gram­mar. In Talmy's terms, information conveyed by the lexicon consists of semantic elements codified in open classes (roots of nouns, verbs, and adjectives), while grammaticali/.ed content is expressed by closed classes (prepositions, pronouns, articles, particles, and all inflectional morphemes). These two components of the

238

linguistic system carry out radically different functions: while closed classes supply the fundamental structure of the language, its structural skeleton, open classes convey elements that are more properly content. This involves a precise qualitative distribution of the information codified in language, because only some items of content can be grammaticalized, while others never are. In particular, the grammaticalized content expressed in the closed classes is topological and non-Euclidean, while content relating to the phenomenological aspect of entities (col­ors, form, size, and absolute dimensions, that is, figurative semes in Greimas' terminology) is never conveyed by closed-class elements. Closed-class elements re­lating to spatial determination (prepositions, deictics such as this, that, here, there, etc.) are neutral in relation to the form, physical appearance, size, and modality of movement of the entities they apply to, and express abstract schematic relations which are topological and relative, never absolute.²⁰ The specific perceptual quali­ties of an object (we could say its exteroceptive semes)—form, size, and general figurative configuration—are, on the other hand, expressed by open-class elements such as nouns and adjectives. For instance, absolute distances are expressed by various systems of numerals, colors by adjectives, and forms by the open class of nouns. Thus, alongside the small number of closed-class terms, such as spatial prepositions, which refer to a relatively few abstract schematic structures of a topological nature, we have a very large number of nouns for specifying the in­numerable forms of objects.²¹

It is interesting to observe that these different fields of specialization of the lexical and grammatical components of the language configure an opposition be­tween perceptual-sensory properties (such as form, size, color, absolute distance, etc.) and structural properties connected to topological schemes that are not based on the sensible properties of the object, but rather on schemata of deep images which may be linked to corporeal structuring.

The search for a semantic basis for linguistic categories undoubtedly charac­terizes the whole cognitive approach to the study of meaning, and also extends to the characterization of the different semantic configurations underlying classes that are formally definable as verbs, nouns, and adjectives. The work of Langacker (1987), Haiman (1985), and Wierzbicka (1985) all moves in this direction; none of them reduce the syntactic to the semantic, but rather all are concerned with the identification of a more adequate principle of semantic motivation which can ac­count for the close connection between form and meaning. To argue that linguis­tic categories have a motivated basis is to claim that they have an ontological foundation which in each case refers to an experiential base, even though this is not necessarily identifiable at the same level. According to Givon (1979, 1984), for example, the main grammatical classes reflect a scale of temporal stability of de­noted phenomena: at one end of the scale we find experiences which are relatively stable in time, linguistically codified as nouns, and at the other end there are experiences of rapid change, that is, events and actions, typically codified as verbs, a linguistic class whose existence is closely dependent on time. Between the two

239

extremes there are experiences of intermediate stability, linguistically represented in the class of adjectives. Givon (1979) also observes that although there are both concrete and abstract nouns, the latter are always derived, generally from verbs. This would seem to suggest that the basic semantic configuration for the nominal class is that which codifies physically determined and spatially delimited entities. In this way, the difference between semantic categories like abstract and concrete derives from distinct ontological bases (a perspective similar to the one I have delineated). Thus it is possible to motivate the different formats of the vari­ous lexical classes by starting from the different experiential saliencies underly­ing them.

The hypothesis of Brandt (1995) is similar; in his view, nominal, verbal, and adjectival classes stabilize different semantic worlds, respectively those of percep­tion, communication, and imagination, which roughly correspond to the fields of physical experience, social experience, and psychological experience. There would thus be a physio-semantics, source of the sub-language of states, expressed by nouns; a socio-semantics, source of the sub-language of events, expressed by verbs; and finally, a psycho-semantics expressed by adjectives and adverbs. Naturally this tripartite division must be taken as a working proposal, but it may constitute a starting point for a more detailed analysis of the semantic-ontological bases of linguistic categories, an area which is as yet largely unexplored.

The more specific problem of the possible extension of the prototype model can be framed in this context. Can the notion of typicality, which has proved useful for the categorial structure of concrete nouns, also be applied to other parts of speech such as verbs and adjectives? The issue is controversial. Some argue that prototype semantics is essentially a nominal semantics; however, studies like Coleman and Kay 1981, Jackendoff 1985, and Fillmore and Atkins 1992 suggest the possibility of extending prototypic analysis to predicates. Pulman (1983) also rec­ognizes a difference in the degree of typicality of actions expressed by verbs, ' and this is confirmed in more recent works on other semantic fields, such as verbs ol perception, which reveal phenomena of differentiated saliency.²⁴ According to Kleiber (1990: 129), however, the intuitive pertinence of a hierarchical structure within nominal and verbal classes is not the same. In his view, one of the dilficul ties of extending the prototype model to non-nominal classes is that often refer­ence is being made to the situation which the predicate denotes rather than CO the prototype of a category. An example of this is the case of adjectives of scale like big, small, good, bad, and so on. Clearly, we cannot talk about a single prototypical meaning for bigot goodifwe do not specify the category of referent to which the adjective is applied each time. A big ant and a big mountain do not have the same prototypical dimensions.

In general, one can agree with Kleiber that there is an intuitive difference in the meaning of" prototypical case when applied to classes such as natural kinds and when applied to adjectives of scale or verbs and the complex scenes that these relate to. I hinted at this problem in chapter 8 when introducing the concepts of

240

frame and scene used by Fillmore. Scenes represent complex situations that de­scribe the context of regularity of a certain action, rather than the central exem­plar of a category.

The idea of a typical scene allows us to grasp a fundamental characteristic of linguistic functioning linked to the effect of distortion that I have already dis­cussed: the relation between our experience of reality and the words with which we speak about it is neither a biunique correspondence nor a precise mirroring. This distorting gap, which cannot constitutively be eliminated, gives rise to a se­ries of procedures of local "readjustment"; processes of analogical extension are one of the most important examples of this, in that they allow the application of a description which is valid for the typical or regular situation to anomalous or deviant situations. Let's consider the meaning of a verb like run. It has been said that the prototypical situation for run is a running man rather than a running crab. Certainly, one could easily object, as Kleiber does, that it is improbable that the categorization of a token of run is made each time on the basis of a compari­son with the prototype of a running man. But the assertion can be reformulated: if "a human being runs" can be considered the typical case of run, it is because this phrase refers to a situation for which we have direct and phenomenologically founded knowledge. Verbs of movement come into the class that I have called natural actions because of an affinity with natural kinds, the linguistic description of which is impossible without referring to the underlying physical-perceptual experience. Run does not only mean "move rapidly" but refers to a particular form of corporeal movement that is understandable only in relation to our upright po­sition and to our having two legs.²⁵

It is, therefore, on the basis of our corporeal experience, the structure of our body, and its vertical position that we know what run means for a human being; starting with this primary meaning we broaden the use of the same expression to describe the movement of a crab or a millipede. A crab or a millipede certainly does not run like a human; this is one of those cases which are continually found in linguistic use where the applicability of a word is extended beyond its typical conditions. In these cases, the distortion of the linguistic description is, so to speak, adjusted, following analogous reasoning procedures: if a crab runs, we imagine that it is moving in a way which has similar features to the way a human runs, even though we know that this similarity is only partial and limited, given the structural difference between the body of a crab and our own.

I believe that many of the equivocations present in discussions like this one would be clarified if, instead of arguing in essentialist terms about the prototype as the best exemplar of the category, the inferential perspective I have indicated were to be adopted. In this perspective, the typical case becomes the hook, and point of departure, for possible inferences, an abductive tool onto which can be grafted an interpretative procedure.

Even if we come up with a new word to indicate the way a crab runs, just as we have trot and gallop for a horse, it would be hard to denominate uniquely the way each animal species rims. Our mnemonic resources would impede i(. Pioce-

The Many Dimensions of Meaning

241

dures of analogical extension are not a choice but a fundamental necessity in order to live with the distortion implicit in linguistic functioning. Thus the intrinsic limitation of language is made up for by its almost limitless flexibility. The num­ber of words at our disposition will always be limited and woefully inadequate to describe the multiplicity of the real: the world contains more things than we will ever be able to name, and our experience is constantly confronted with the limits of our words. But at the same time our language always seems able to transgress its own limits, and also our own.

NOTES