Lexicography of the Verb and Cognitive Linguistics: A case study of Afshandan

The work of a linguist and a lexicographer in the field of lexical semantics has the largest common border. In the process of lexicography, a lexicographer faces one of the main issues in lexical semantics: polysemy. When a lexicographer is trying to define a polysemous word for a monolingual dictionary, he/she has to undertake a special task, so-called: sense discrimination, that means he/she has to make a distinction between various meanings of that word. Lexicographical practice in short appears to be in accordance with the lexicological observation that the distinction between meanings need not to be clear-cut. This has been a controversial problem in both disciplines. In order to provide some argumentations to the problem, this research is conducted with the help of the descriptive tools that cognitive linguistics offers, namely: the theory of Semantic Networks (Norvig & Lakoff, 1987), Frame Semantic (Fillmore, 1982) and Construction Grammar (Goldberg, 1995; Nemoto, 2005).  This study is conducted on the lexical category of “verb”, which has semantic complexity, and to this end, the Persian motion verb of Afshandan has been chosen as a case study. The data for this research have been extracted from the corpus of the Academy of Persian Language and Literature, which includes about 500 sentences and phrases containing this verb.
The synonymy problem: By examining the syntactic and semantic distribution of arguments of some so-called synonymous verbs, which are used in the definition of the verb afʃãndan in Persian dictionaries, it is intended to demonstrate the usefulness of Fillmore’s (1982) Frame Semantics for describing verbal argument realization patterns across these near-synonymous verbs. This section addresses the issue of describing the similarities and differences exhibited by synonymous verbs which are routinely used for defining the verb afʃãndan, namely: rixtan (to pour), pãʃidan (to spray), parãkandan (to scatter) and andãxtan (to drop). The most significant motivation inside the language for making a sense discrimination is the existence of these synonymous verbs which are almost equal to each sense of afʃãndan and another motivation for the description of the frame verb is profiling the “destination” in the lexical semantics of these verbs. We see in lexical semantics of the verbs like rixtan, pãʃidan and andaxtan as apposed to parãkandan, the “destination” is profiled and it is hidden in the meaning of the verb. The motivation outside the language is the categorization of patients, which indicates the meaningful difference between the properties of the patient role of the verb andãxtan in comparison with rixtan, pãʃidan and parãkandan. The patient of the third one would be a massive thing that does not have a potentiality for turning into tiny fragments by a light force. The result of this section is also checked with the definition of the equal English verb afʃãndan: “to scatter” and it reveals many similarities between both Persian and English verbs’ frame elements in the process of sense discrimination.
So, profiling the “destination” in lexical semantics and categorization of the objects (patients) are two reasons, which justify discrimination in the process of defining a transitive motion verb like afʃãndan.
The hyponymy problem: The most significant contribution that Construction Grammar can make to lexicography is the information about the syntactic behavior of words that is or could be included in a dictionary. In the current study, by interacting the verb afʃãndan with the preposition construction “az” which makes the “source” in the verb’s frame semantic elements, profiled, we face a new sense which arises from the following construction: “object+ az+ source (place/ surface)+ verb (afʃãndan)”. This new sense would mean like “wiping”. Now there is a motivation inside the language, e.g. in our corpse, that is, most of the words used as objects of the construction “az …afʃãndan” could be categorized in the category named “dust” and this sense is like “dusting” in English.  If we can consider dusting as a kind of wiping, it seems we face a problem; we have named it “hyponymy problem” in sense discrimination process. It is argued that in favor of evaluating the frequency rule in the corpse-oriented studies and cognitive linguistics, we can consider the incorporated verb “gard (dust) afʃãndan (wiping)” as a new sense that means “dusting”. What if we consider “gard giri” (dusting) under hyponymy of “roobidan” (wiping)? In another sense discrimination by profiling the “intention of agent” and “result of the act” in addition to a new frame (shaking the source, e.g. place/surface), it seems that we would have a new sense. At the same time, there is a motivation inside the language, which is the existence of the synonymous verb “takãndan” (shaking the surface of something in order to remove the tiny things from it) equal to this sense of afʃãndan. Again there is another hyponymy problem because we have some incorporated form of the verb as “dast (hand) afʃãndan”, “dãman (skirt) afʃãndan” etc., in which we have just a sense of “shaking” for afʃãndan. So, can we consider “takãndan” under the hyponymy of “takãn dadan”?
It seems that a solution for this problem is considering information about “constructions” like incorporated and prepositional constructions, as important information that must be included in dictionaries.
The existence of lexical network: Not all different meanings exist in isolation; they are related in various ways to the central sense and to each other’s. The existence of different types of connectivity between different meanings in the lexical network of the polysemous verb could be important for lexicography because a lexicographer can avoid wrong analogy between the possible sense of verbs and the existence of a near-synonymous verb for it in the language. For example, creating a sense of “afrooxtan (to fire)” for afʃãndan when it used with the noun like atash (fire) or doozax (hell) is an over-generalization because we cannot connect this meaning to the central or the other meanings.
 
Figure 1. Semantic network of the verb “afʃãndan”
Lexicography, as a highly specialized domain with general and specific readers, is greatly influenced by linguistics. This research represents some aspects of the Cognitive Linguistics theory throughwhich the corpus data can be identified and analyzed in a more systematic and less subjective way.