-
Chapter 3. Usage and Cognition
- Slavica Publishers
- Chapter
- Additional Information
56 3. USAGE AND COGNITION (2) underlying form # р э к a´ #NOM.SG # р э´ к и #NOM.PL consonant softening # р’э к а´ # р’э´ к’и ikanje # р’и к а´ does not apply phonetic #[р’и к а´] #[р’э´ к’и] This is how generative phonology accounts for the relationship be-‐‑ tween [и] in река´ and [э] in ре´ки. An underlying (abstract) form of a word is set up and all the actually occurring forms of that word are generated from the underlying form. 3.2. Underlying Forms Versus Usage The generative approach has much to recommend it, not the least be-‐‑ ing its systematic methodology. We do not wish to discard what is useful about this method, but generative phonology has some im-‐‑ portant drawbacks. First, the underlying form—such as #рэка´#—does not occur in Russian in any word. Paired consonants are always soft before front vowels, and we know that the sound [э] can occur only under stress. Therefore the proposed underlying form does not exist in reality. Generativists answer that the underlying form is a handy ab-‐‑ straction that generates, with proper rules, all forms that do exist. Our point is that as an abstraction it represents no actually occurring form. In a usage-‐‑based approach, such as the one presented in this text, variants such as [р’ика´] and [р’э´к’и] are learned by experience and exist side by side in the speaker’s mind. There is no need to generate either one. The speaker can choose whichever one is needed in a given context. Generalizations (i.e., rules) which relate alternations emerge in a speaker’s mind based on actual linguistic input, i.e., based on what the speaker actually says and hears every day. In a usage-‐‑based pro-‐‑ gram, the structure of the language is overt, it is not derived from a deeper (unverifiable) abstraction. It therefore seems reasonable to as-‐‑ sert that a usage based description, such as the one presented in this text, renders a better picture of reality than the generative-‐‑based description. [3.17.150.89] Project MUSE (2024-04-24 22:37 GMT) 3.3. SCHEMAS 57 3.3. Schemas Throughout this text we use a simple tool, the schema, to describe re-‐‑ lationships between specific linguistic tokens and the meaningful cate-‐‑ gories to which they belong. For the various forms of the word река´, for example, we write the schema: (3) ‘RIVER’ [р’ика´] [р’э´к’и] [р’ику´] … (The three dots above indicate there are other forms that could fit in here, too. We’re just not bothering to list all the forms of this word.) Note that in this schema only actually-‐‑occurring forms are part of the description. When children learning Russian first hear these words (in any case form) they store them away in memory as pronounced, hav-‐‑ ing learned what their referent is. Later they misspell the word in school, assuming that и stands for all instances of [и], because they have heard [р’ика´] (or another singular form) much more often than the plural [р’э´к’и]. (Shteinfeldt reports this word is used 84% in the singular, 16% in the plural.) The second reason why we find the generative approach untenable is that it implies that whenever speakers want to use, say, the genitive form of the word, they apply one or more phonological rules to an abstraction to come up with the proper pronunciation. We think this is unlikely. Even more unlikely is the flip side of that coin. If speakers hear [р’ик’а´], they must apply some rule to degenerate this word to its basic abstract form [рэка´]. The problem with this is that, as discussed in chapter two, phonetic [и] may be associated with /э/, /и/, /a/, or /o/! No generative rule could pick the proper association without looking at other forms of the word, which would essentially make it a usage-‐‑ based approach. In cognitive grammar we deal with “abstractions” (such as pho-‐‑ nemes /a/, /o/, /т/, etc.), but only as names of categories of sounds. Thus the phoneme /a/ is an abstraction in that it represents all the pos-‐‑ sible pronunciations of a low back vowel in Russian, but it also repre-‐‑ sents the pronunciation of [a], a low back vowel. The same is true with meaning. The “abstraction” /RIVER/ refers to whatever rivers the 58 3. USAGE AND COGNITION speaker knows about and may even represent a prototypical river, but it also relates to the Volga or some other specific actually occurring river. The abstraction is nothing more than a useful category name that comes from the actually encountered instances. (4) /category/ [instance1] [instance2] [instance3] Where the category name is one of the instances and stands for the rest. Why is any of this important? Aren’t...