|
Article Excerpt
This paper presents a cognitive linguistic analysis of CONSTRUCTIONAL POLYSEMY manifested by Tagalog ka- and its variants. We propose that ka- is a METONYMIC OPERATOR that functions to evoke and mark a SEMANTIC PARTIAL, which we define as the conventionalized profiling of an element that is selected of abstracted from the conceptual base evoked by a linguistic root of stem; subsumed by the partial schema are the categories of INDIVIDUAT1ON and ABSTRACT QUAL1TY. When the cognitive linguistic analysis of ka- as a semantic partial is applied to more complex constructions, it reveals motivations that are missed by other approaches. Our findings are compared to Panther and Thornburg's (2001, 2002) study of METONYMY in the English nominalizer -er, which exhibits a roughly similar range of meanings. We support their conclusion that lexicon and grammar form a continuum of symbolic units.*
INTRODUCTION. (1) In the Panganiban (1972:875) Tagalog dictionary-thesaurus under entries for the unglossed word root sama, we find the following glosses of the derived term kasama: 'companion', 'part of a mixture of whole', 'husband of wife', and 'included'. The construction kasamahan is glossed 'fellows of companions in a group'. Since sama itself might be glossed as 'together' or 'included', perhaps some of the variance is contributed by alternative construals evoked by the prefix ka- and the circumfix ka- -- -an. On first inspection, ka- appears to be a nominalizer, similar in function to the English derivational suffix -er/-or, which is typically regarded as an agent or instrument nominalizer (e.g. rescuer, screw-driver) (See Payne 1997:226, 228). But detailed study of -er usages has proven that morpheme to be more complex. Panther and Thornburg (2001) presented more than a dozen variant meanings centered around the prototype 'human Agent who performs an action or engages in an activity to the degree that doing so defines a primary occupation' (2001:154). Many, probably most, of the variants are best explained as motivated by metonymy. The term sleeper, for example, 'may also be used to denote (a child's) clothing designed for sleeping, the specially designated car on a train for sleeping, and a drug/event that caused sleep' (2001:191-92).
We will show that the category structure of Tagalog ka- has much in common with English -er as described by Panther and Thornburg, even though the overlap in the semantic domains of the two affixes is minimal. However, for ka- we propose that a single schema of partial subsumes two important subschemas: an individuation schema, which is typically nominal, and an abstract quality schema, which is typically adjectival. These in turn have important subcategories as shown in Figure 2, which uses conventions of heuristic diagramming developed for cognitive linguistics by Langacker (1987). The conventions are illustrated in Figure 1. The dotted lines to the blowup of the D box in Figure 1 have no theoretical significance. The schema and subschemas of ka- are attributed not to the prefix alone but to constructions with various classes of linguistic roots and stems. Another important sense of ka-, the RECENT COMPLETIVE, is verbal, thus completing for this prefix the triad of noun, adjective, and verb that de Guzman (1996b) argued is found in Tagalog. We treat the recent completive ka- forms as lying outside the partial schema.
[FIGURES 1-2 OMITTED]
In the English language literature, analyses of ka- seem to begin with Bloomfield (1917:265-98), who devoted several pages to its grammar, approximating our own strategy of first determining its senses when it is the only affix in a lexeme, and secondly its senses in complex lexical constructions with multiple affixes. He emphasized its relational meaning, but he did not attempt to relate the various senses of the prefix to one another nor to explain the senses produced in compound affixes by reference to the senses of ka- as the only prefix in a lexical construction. In other words, he did not attempt a unified account. Blake (1925) saw individuation as important to understanding ka- and place as important to the meaning of the circumfix ka- -- -an. Lopez (1941) provided a list of terms similar to those of Bloomfield and Blake.
In the most recent comprehensive grammatical description of Tagalog, Schachter and Otanes (1972) collated constructions according to traditional grammatical categories. Their structural approach seems to assume that function morphemes such as the verbal affixes evoke no consistent meanings. Thus, ka- is never singled out for examination as a unit in its own right but is found scattered throughout their grammar in connection with a variety of constructions. The few examples that they discussed explicitly will be taken up in the appropriate following sections. (2)
Terms that have ka- as their only affix occur with relatively low frequency. Only a half-dozen distinct terms appear in over six hundred lines of text from transcripts of three interviews that were conducted for a previous study. (3) But ka- is important to Tagalog grammar because it appears very frequently in many compound prefix constructions ranging over such diverse semantic entities as verbal mode and aspect, superlatives, and derived generic nouns. Analysis of the category structure reveals semantic motivations governing constructions with ka- that are missed by approaches that assume a more arbitrary relation of form and meaning or that would simply classify ka- as a nominalizer. We find that Tagalog lexical constructions are less arbitrary than they are made to appear in the linguistic literature. In this paper we collate glosses and usages of constructions containing ka- in order to determine its conventional meanings and outline its category structure. The assumption is that the prefix is a complex category with one or more central schemas, each of which may serve as the origin of a radial structure based on relations of similarity and contiguity, in which we include, respectively, metaphor and metonymy (Lakoff 1987, Langacker 1987, 1991, Palmer 1996, 2003a). In Figure 1, boxes represent entrenched concepts. Relations of schematicity (the inverse of INSTANTIATION) are depicted with solid lines. The hierarchy of schematicity implies that categories and terms are fully sanctioned by the immediately superior category. Relations of similarity (the inverse of variation) are depicted with arrows having dotted fines pointing from reference concepts to their variants. Variants (extensions) are only partially sanctioned by the reference concept. In fact, all subcategories of a schema are related by similarity, and all variants on a reference concept share a schema with it. To avoid cluttering the figures, we have omitted the arrows with dotted lines. Arrows with double lines represent the relation of metonymy.
In our experience, most affixes exhibit the kind of polysemy described above. One might attempt to go further and identify prototypes on the basis of concreteness of conceptualization or frequency of usage, but we have refrained from doing so because we lack a sufficiently large sample of terms, across representative discourse contexts, to enable a meaningful determination of frequency. Also it is not obvious which concepts Tagalog speakers might regard as most concrete. For demonstrations of the radial structure of spatial prefixes, see Occhi, Palmer, and Ogawa (1993) and Ogawa and Palmer (1999). Prior to these studies, Casad and Langacker (1985) and Langacker (1991) showed that variant schemas govern the polysemy of Cora spatial suffixes. This work is closely related to work on the polysemy of prepositions, such as the studies of Brugman (1988), Sandra and Rice (1995), and Rice (2003).
Our analysis has another important dimension. We argue that the polysemy of ka- originates in conceptual metonymy, which we believe is a more pervasive cognitive process than conceptual metaphor. To understand the analysis by metonymy, it is useful to review Langacker's (1987:183-9, 2000:6-7) distinction between PROFILE and BASE, one of several distinctions in his theory that involve the relative prominence of conceptual entities. The profile of a lexeme or construction is the concept that it designates. The base (or SCOPE of predication) is the full conceptualization needed to interpret the profile. It is the ground against which the profile is the figure. One can think of the conceptual base as a scene or scenario evoked by a term in a particular discourse context. In other words, the meaning of a term includes some conception of the context of usage. For example, the term spoke in the context of wheels (rather than speech), makes sense only against a background conceptualization of an entire wheel with spokes. The term spoke designates and, therefore, profiles a certain part of the wheel that, in the speaker's thoughts at the time of utterance, has greater prominence than other parts such as the rim or the hub. Similarly, a kinship term, such as aunt, makes sense only against a conceptualization of a kinship network that includes at a minimum a reference individual, his/her parent, and the parent's female sibling.
Now imagine an affix that combines with various linguistic roots to evoke profiles that are always just a part of or selection from the full conceptualization evoked by the linguistic root. The part that is profiled for each such construction is determined by convention; it may be a part of the usual profile of the uncombined root or a part of the root's conceptual base. This is what we mean by a semantic partial. In fact, a partial can be either a metonymic target or an abstracted element. This is a kind of conceptual metonymy that is triggered by the affix according to convention. In such a process, one can expect that roots belonging to various semantic categories will be the sources of particular types of constructional profiles. Additionally, it seems likely that derivational categories will correspond to semantic categories and subcategories of roots. In this paper we will attempt to illustrate why ka- has led us to define this new category of grammar--the semantic partial--that shows promise of having wider application.
The various categories of ka- constructions are actually subsumed by two schemas, which we have termed PARTIAL and EXPERIENTIAL (Figure 2). The domains subsumed by these schemas are complexly interconnected. With the partial schema, the individuation subschema is ambiguous between ONE OF A COLLECTION and PART OF MASS. The degree to which a collection of persons may be conventionally construed as merged into a single conceptual mass is presently indeterminate. We left the question open by using the neutral label individuation. (4)
Like the partial domain, the experiential domain is hypothetical. The notion of abstract quality emerges from the universal cognitive process of schematization. Abstract quality is a meta-schema: It is what is common to a set of abstractions. Since abstraction is a cognitive process, it seems reasonable to place it in the experiential domain, which includes cognitive as well as emotional experience. The subcategories COMPARISON OF EQUALITY and INTENSE EXPERIENCE both involve the registration of some element of psychological experience as does the notion of SHARED EXPERIENCE. The recent completive, not included in the partial schema, is an aspect of experience located on the realis margin of the incipient or inchoative moment. This is the margin that is proximate to the temporal reference point, often the time of speaking. These are the considerations that led us to posit the experiential schema.
Ka- attaches to roots with quite diverse semantics to derive new lexemes and stems. For example, kaputol 'one piece' (
For some of the semantic classes of ka-, it can be said that they are categorized by the classes of their lexical roots. For example, only psych-predicates make up the set of roots in the semantic class of intense experience. Roots in this class undergo complete reduplication (R2) of the root or stem, of the first three segments of these. Roots for 'eating' and 'drinking' have historically provided semantic bases for terms denoting companions as with the obsolete kaatay 'companion in eating liver' and kanubig 'companion in drinking water' and the contemporary kainuman 'drinking buddy'.
For other semantic classes of ka-, the roots may variously evoke processes, things, or states. For example, the roots of the PARTICIPANT IN SHARED EXPERIENCE category include those with glosses 'cut' (process), 'race of human beings' (thing), 'companionship' (thing), and 'intimate' (state). These are not only very different categories but also different levels of abstraction. These observations, and further complications to be discussed later, suggest that each of the semantic classes of ka- has its own construction or constellation of constructions: Each is a kaputol, one piece cut from a semantic matrix, and the collected meanings of ka-constitute a semantic archipelago, a ka-pulu-ang ka-hulug-an (< pulo 'island', hulog 'meaning'). Apparently, in this construction with ka-, the locative -an suffix (-ang < -an + na) pluralizes and rederives a place or thing as a reaggregation of previously individuated items. It is, thus, a beautiful example of linguistic iconicity in which there is first a cutting off with ka- in initial position, followed by a reassembly with -an in final position. Jean-Paul Potet (personal communication, June 2003) also noted the form sangkapuluan 'the whole archipelago'. The notion of an aggregate can be seen also in one of the terms cited earlier: kasamahan 'fellows or companions in a group'.
We have described the outlines of a category network and something...
|
