Linguistic Methods, 3 October 2006 4th session: Work with informants–language elicitation

Linguistic Methods, 3 October 2006 4th session: Work with informants–language elicitation

• Karl Erland Gadelii, Linguistics Department, Gothenburg University ([e-mail])

Note

• This presentation was inspired by Jouni Maho: “Datainsamling genom språkelicitering” ( = chapter 5 in the manuscript Lingvistisk Metod, ed. Jens Allwood, Linguistics Department, Gothenburg University).

Introduction

• Language elicitation = one or more informants (normally native speakers of the language under investigation) are made to produce linguistic data. The procedure is also called “informant work” or “linguistic interview”, and contrasts with data collection from written sources. There is a certain overlap between language elicitation and methods such as observation, interview and questionnaires, which will be treated by Sally Boyd in the next session of this course.

Introduction II

• Language elicitation is often (but not necessarily exclusively) carried out when investigating “exotic” languages that the linguist does not speak him/herself. However, the classical chomskyan way of consulting one’s own linguistic intuitions concerning one’s mother tongue could also be seen as a kind of elicitation.

Introduction III

• Language elicitation has traditionally been used within North American field linguistics, but can also be used in more armchair linguistics-type environments with exiled speakers of “exotic” languages. Nowadays, language technology plays an important role in language elicitation.

Why linguistic descriptions?

• documentation of endangered language and/or culture
• development of orthography
• production of textbooks, manuals, etc.
• compilation of dictionaries and grammars
• investigation of a certain linguistic phenomenon
• translation
• constructing terminological databases
• language cultivation and planning, language standardization
• educational aims
• language-internal theoretical relevance: parametric variation within Universal grammar, typological study

Partial vs. complete language descriptions

• phonetics/phonology
• morphology
• syntax
• semantics
• pragmatics
• sounds, syllables
• morphemes
• phrases (in the linguistic sense), sentences
• oral and written texts
• communicative behaviour

• (more on this soon)

Two important questions

• What could the linguist expect to encounter during elicitation (at all levels)?
• To what extent are Indo-European languages (which are often those the linguist speaks) representative of the languages of the world?

”Exotic phenomena for the Western linguist

• complex phonemes
• different tyes of phonation (creaky voice, breathy voice, etc.)
• tones (lexical and grammatical)
• floating tones
• ergative-absolutive case instead of nominative-accusative
• serial verb-constructions (“he threw the bottle break” –> ”he smashed the bottle”)
• predicate clefts (”dislike I dislike pizza” (=”I really dislike pizza”)
• fourth person (different forms for he/she/it depending on the location and relevance of the referent)
• unalienable possessives (“Mary’s books” is expressed differently from “Mary’s eyes”)

Exotic phenomena II

• relevance of stative vs. dynamic feature of verbs (cf. however English “he is reading the book” vs. “*he is knowing his lesson”
• relevance of proximity (temporal or among referents)
• basic word orders other than SVO
• complex person marking on verbs (extending beyond person and number)
• extensive marking of TMA and valency
• grammatical hierarchies based on animacy
• ideophones
• special narrative forms of the verb

How do the linguist and the informant communicate?

• The linguist (L) knows the language of the informant (I) (advantageous but unusual, would in extreme cases mean that no informant is needed!)

• L and I have a language in common which is not the language under investigation (practical but presupposes equal competences and translatability)

ctd.

• Via interpreter (good but time-consuming, in addition the interpreter may be as good an informant as I!)

• L and I have no means of communicating verbally, but use gestures, signs, non-verbal communication, drawing in the sand (inefficient, restricted)

Elicitation techniques

• Direct elicitation: pointing (good but restricted)

• Simple elicitation: I translates (presupposes that I knows L’s language and that translatability obtains)

• Corrective elicitation: L suggests words and expressions in the language under investigation and I corrects (nice method but demands skilful and inventive L)

Elicitation techniques II

• Elicitation by paraphrase: L asks I to say the same thing but in another way (could be useful but there is a danger that I tries to figure out what L “wants to hear”)

• Paradigmatic elicitation: (I sleep, you sleep, he/she/it sleeps; I sleep, I slept, I will sleep,… (exhaustive but boring)

• Targeted elicitation: L observes recurrent phenomena in the language investigated, formulates hypothesis and constructs various test sentences designed to elicit more of the phenomenon in question (interrogatives, negation, semantic fields (colours, kinship terms), conditionals (cf. Boyd chaper 4) (could be very useful in case the input is aptly designed)

Approaches

• Bottom-up method: start by eliciting individual words in order to get a grip of the phonology of the language

• Top-down: try to identify smaller structures in oral or written texts (more difficult)

• Compromise : work with “kernel sentences”, containing both nominal and pronominal arguments.

Transcription

• Major problem in field linguistics, largely unnoticed in armchair linguistics. (To be kept apart from the question of how to render spoken language in writing in an already standardized language, such as e.g. Swedish).

Vaux & Cooper (1999:28):

• The International Phonetic Alphabet (IPA)
• Official writing systems already used for the language in question (if any)
• Orthography based on dominating language in the area in question
• Established system used by linguists who have previously worked on the language in question
• Newly-invented orthography (cf. Cree and Inuktitut)

Linguistic levels: Words

• Nouns are the easiest words to elicit: natural phenomena, body parts, plants, flowers, trees, animals, food, household utensils, cooking methods, clothes, jewelry, tools, artefacts.

• As pointed out above, lexical items give info about the phonological system of the language at word level.

Phrases

• NP, VP, PP, AP, AdvP. Often interesting contrast between different attributes: the roof of the house, the house’s roof, the black roof. Gives information about morphosyntactic categories, certain word orde facts, gender/noun class systems, agreement/concord, local cases, phonological and tonological sandhi.

Sentences

• Kernel sentences practical but problematic since lots of TMA-information and similar phenomena may get lost. Also a bit monotonous and counterintuitive for the informant if the individual sentences do not form a whole together. Next steps: coordination and subordination of sentences. Gives info about word order, inflectional morphology, TMA, conjunctions and subjunctions, prosody at sentence level, yet more cases of sandhi.

Texts

• The least boring elicitation method. Could comprise mythological and religious stories as well as everyday narrations about hunting or how to build a house. TMA markers naturally fall into place and certain discourse-induced items may suddenly appear. In addition, a lot of useful pragmatic information is gathered.

Characteristics of a good language description

• (cf. the sessions on the Philosophy of Science)

• “natural” (describes the vernacular)

• coherent (does not contain contradictions)

• complete (covers morphosyntax, lexicon and pragmatics)

To keep in mind:

• The linguistic theory embraced by the linguist and
• his/her linguistic prejudices will influence how the
• language in question is described

Before closing…

• "In linguistic communities where until

• recently books and electronic media have

• been absent, story-telling is an art, and skilful

• narrators enjoy the highest prestige.”

Bibliography

• Bouquiaux, Luc & Jacqueline M.C. Thomas. 1992 (original version in French, 1971). Studying and describing unwritten languages. Dallas: Summer Institute of Linguistics.

• Gippert, Jost, Nikolaus P. Himmelmann & Ulrike Mosel. 2006. Essentials of language documentation. Berlin/New York: Mouton de Gruyter.

• Grenoble, Lenore A. & Lindsay J. Whaley. 2005. Saving languages. An introduction to language revitalization. Cambridge: Cambridge University Press.

Bibliography (ctd.)

• Maho, Jouni. 1996. Datainsamling genom språkelicitering. In Lingvistisk metod, Inst. f. lingvistik, Göteborgs Universitet.

• Payne, Thomas E. 2006. Exploring language structure. A student’s guide. Cambridge: Cambridge University Press.

• Samarin, William J. 1967. Field linguistics. New York: Holt, Rinehart & Winston.

• Vaux, Bert & Justin Cooper. 1999. Introduction to linguistic field methods. München/Newcastle: LINCOM EUROPA.

• Wray, A., Trott, K. & Bloomer, A. 1998. Projects in Linguistics. A Practical Guide to Researching Language. London: Arnold

Exercise

• 1. Criticize the dichotomy partial versus complete language description.

• 2. Discuss who should be chosen as informant(s) for a language description of Swedish.

• 3. Explain briefly different types of elicitation.

• 4. How does the language-eliciting linguist affect the data elicited?