The efficiency theory of asymmetric coding: An overview

21 октября 2019 года
аудитория П-7
начало в 18:00


Martin Haspelmath

Professor Dr. at Max Planck Institute for the Science of Human History,
Honorary professor of linguistics at the Leipzig University



In this talk, I give a short overview of the efficiency theory of asymmetric coding, which explains a large number of grammatical patterns on the basis of a few simple principles that are well motivated by other observations (going far beyond grammar, and even beyond language). The term asymmetric coding refers to a situation in which two contrasting grammatical patterns differ only in that one of the patterns shows shorter coding and the other one shows longer coding. In most cases, shorter coding means zero-coding, and longer coding means an overt marker contrasting with zero, as seen in (1a-b).

(1) asymmetric coding of two contrasting grammatical meanings

       a. singular – plural                                            (book-Ø/book-s)
       b. present tense – future tense                         (Ø go/will go)
       c. allative – ablative                                          (to/from)

Under asymmetric coding, I subsume two kinds of situations: Cases where two contrasting grammatical meanings are coded differently in an asymmetric way, as in (1a-c), and cases where one and the same grammatical meaning (such as patient, or imperative) is coded differently in an asymmetric way, depending on the grammatical context or the lexical subclass, as in (2a-d).

(2) asymmetric coding differentiating between two contexts (a-b), or two subclasses (c-d)

       a. patient:
          indefinite – definite           (Hebrew Ø sefer/et ha-sefer)

       b. imperative:  
           2nd person – 3rd person    (English Ø come!/let her come!)      

       c. reflexive:
          extroverted – introverted   (English shaved Ø/hated himself)

       d. adpossessive:
           inalienable – alienable       (Old Italian moglia-ma/terra mia)                                                                                                  

Asymmetric grammatical coding is found very widely throughout the world’s languages, and in cross-linguistically recurrent ways. I propose that all of these patterns are reflections of a general causal force of efficient coding: Frequently used and therefore predictable grammatical meanings and constructions get short (or zero) coding, and rarely used meanings and constructions get longer coding. I will suggest that this theory fares better at explaining a range of famous phenomena (split ergativity, passive marking, singulative marking, word-class distinctions, and all kinds of phenomena subsumed under "markedness") than other theories.