American structural linguistics

Structural linguistics in Europe was partly concerned with meaning and interpretation but in N. America Franz Boas and Leonard Bloomfield took a more descriptive/ positivist stance. They also reversed de Saussure's emphasis on the creative aspect of everyday language. American linguistics began as an offshoot of anthropology and was motivated by the urgency of studying and preserving the American Indian languages which were fast dying out ... Boas's student Edward Sapir came closer to European interpretivism - see sec. .

Leonard Bloomfield 1887-1949 Early work Introduction to the Study of Language, 1914 was influenced by Wundt's mentalism. In his later Language, 1935 he adopted a strict and explicit behaviouristic stance. (See story of Jack, Jill and apple tree in Lyons, p.32 - language as a substitute for other, non-symbolic, behaviour)

• Led to development of an empiricist methodology for study of language.

• ``The only useful generalizations about language are inductive generalizations.''

• Since precise definition of most words was impossible to give, semantics was assumed to be outside the scope of linguistics.

• Distinction between descriptive and prescriptive (i.e. normative) grammar - the emphasis was on the former.

In psychology Behaviorism (1920s-1950s) rejected the focus on mental processes and stressed measurement based on objective behaviour.

How does experience (reward and punishment) shape behaviour?

B. F. Skinner (1904-1990) Children learn language through shaping (correction of speech errors).

The associative chain theory: each word in a sentence serves as a stimulus for the next word, and the entire sentence is produced left to right.

Approach to studying a language:

1. Discover the basic sound elements (phonemes) of the language.

2. Discover the rules by which these phonemes are combined to form sentences.

Post-Bloomfieldians aimed at more objectivity (even mechanisation) in phonological and morphological analysis. The ``sound spectrograph'' was invented at Bell Labs and made public in 1945; it analysed speech samples for physical properties. Machine translation projects were started in the early 1950s and machine parsers appeared in the late 1950s.

At the same time the post-Hullians in Psychology worked on mediation theories of language. Psychological reality of linguistic constructs (eg. phonemes) was studied by some. Subjects' judgements of similarity between speech sounds were factor analysed to identify space of speech sounds. Analyses were also done to ``measure meaning'' (Osgood, 1957).

The first half of the 20th C saw a confluence of historical, comparative and structural descriptions of language. Phonology and morphology was well developed by this time but not so syntactic analysis. Phonology served as a model for grammar. The idea was to build from word level (morphology) to syntax and thence to semantics.