Discourse, Fall 1997 9/18: "Toward a Taxonomy of Given-New Information" by Ellen F. Prince
She begins by describing the general assumption that the purpose of language is to convey information. This may be social or ritual as well as "objective" (what we more typically think of as information content.) But not all objective information is of the same type, in particular, it is commonly understood that some information is older (or newer) than other. This is essentially the "Given-New" idea that is well established in linguistics although there is no common or uniform definition of what these terms entail.
In the rest of her article, she reviews 3 different concepts of Given-New, examines one in greater detail, analyzes some examples of discourse, and points to questions for further research.
1) Predictablity/Recoverabilty:
This version defines Givenness thusly:
The speaker assumes that the hearer CAN PREDICT OR COULD HAVE PREDICTED that a PARTICULAR LINGUISTIC ITEM will or would occur in a particular position WITHIN A SENTENCE. (226)
She attributes this use, more or less, to Kuno and Halliday, though for Kuno the notion of recoverabilty is important, for Halliday, the predictabilty is marked intonationally. This difference results in different interpretations of utterances in some cases. She suggests that both recoverability and intonation provide important information and that the problematic cases can make sense when combined, with the addition of the Parallelism Principle:
A speaker assuumes that the hearer will predict, unless there is evidence to the contrary, that (a proper part of) a new (conjoined?) construction will be parallel/equivalent in some semantic/pragmatic ways to the one just processed. (228)
2) Saliency:
Here givenness is defined as follows:
The speaker assumes that the hearer has or could appropriately have some particualr thing/entity/... in his/her CONSCIOUSNESS at the time of hearing the utterance. (228)
Known and unknown can be equally "new" in a discourse and a "given" can be either explicitly introduced or physically present.
3) Shared Knowledge:
The "givenness" of shared knowledge is described as follows:
The speaker assumed that the hearer "knows," assumes or can infer a particular thing (but is not necessarily thinking about it).
She finds this in Clark and Haviland, and also in Kuno's anaphoric, and it can include information from inference (bridging) as well as explicit statements.
These three understandings of givenness are different, but not mutually independent.
After this outline of the various concepts of givenness, she turns her attention to "so-called 'SHARED KNOWLEDGE'"
The term itself is problematic as it assumes some omniscient observer viewpoint to identify the region of sharedness, when in normal discourse this is negotiated much differently. As already explained, "givenness" is a poor subsitute. The term she prefers is "ASSUMED FAMILIARITY".
The problem is to understand what kinds of speaker assumptions about the hearer have a bearing on what is said--with a specific focus on the linguistic productions. A solution will have 3 parts: a taxonomy of linguistic forms; a taxonomy of values of Assumed Familiarity; and an account of the correlation between the two.
She suggests a solution by means of an analogy with recipes. With respect to Assumed Familiarity, it is not clear whether is is binary or some sort of a continuum. She suggests that a text is like a recipe, insofar as it is a set of instructions about how to produce discourse. Discourse entities are then analogous to ingredients: a new discourse entity is an ingredient placed on the counter.
It can be brand-new (not previously available) or Unused (commonly on hand). Brand-new entities can be anchored or not to other discourse elements. Elements that are "already on the counter" are not new; they can be evoked textually (explicit reference) or situationally (physical presence). And evoked entities can also be inferrables.
She then works through two examples of naturally occuring discourse (one oral, one written) looking for these kinds of distinctions.
Some of her key points:
in the oral conversation:
--nearly all subjects are evoked, less than half the nonsubjects are; one sixth of the nonsubjects, but none of the subjects are new; and inferrables are one third of the nonsubjects, but relatively few of the subjects
--inferables are mostly culture based, relying on Stereotypic Assumptions
--the Conservation Principle states that hearers do not like to make new entities when old ones will do and co-operative speakers will make maximal use of old entities
--the use of Brand-new Anchored entities seems to be characteristic of informal discourse.
--in informal conversational discourse the subject position is reserved for relatively more familiar NPs
in the written text:
--there is much more metalinguistic inferencing than in the oral one.
--there are many more highly abstract entities
--there is a notable blurring between the categories of Unused and Inferrable which does not arise in the oral narrative.
--the cultural assumptions required are more complex and abstract
Areas for further study include:
does her taxonomy of Assumed Familiarity hold? Once this is established, it will help with the following issue
there is a need for a deeper and clearer account of the correlation between form and understanding in natural language
and both of these could help in the study of what reading and writing entail beyond coding and decoding.