Activation states; reference points

statuses are generally more transient, and have to do with whether or not a concept is in active consciousness. These concepts are further discussed in the following sections.

2.4.1 Activation states; reference points

At any given time in the comprehension of a discourse, concepts are in one of three states of activation in relation to the addressee: active, semiactive or inactive see Dooley and Levinsohn 2001, ch. 10 and references there; Chafe 1987, 1996 and 1994, ch. 6 and 8; Lambrecht 1994, §3.3. • An ACTIVE or GIVEN CONCEPT is one which is in the addressee’s consciousness, hence also in her current mental representation. It is like a light which is turned on or what we see in the active focus of vision “foveal vision,” Chafe 1990:89. Commonly, a speaker activates a concept by what he says, so that subsequent references use minimal coding §2.4.3: This is a carburetor. It mixes fuel with air. Concepts can be activated in dialog: A: What does a carburetor do? B: It mixes fuel with air. Concepts can also be activated extralinguistically, as when one points ostensively to something or notices that the addressee is already conscious of something. Thus, a car salesman may come up to customer and point to a car or simply notice that the customer is looking at a car: It’s a beauty, isn’t it? This shows that activation states are part of general cognition and broader than language per se. • A SEMIACTIVE CONCEPT is one that is in the addressee’s current mental representation, at least as a plausible role or slot, but is only partially than fully active. It is like a dim light or like something in peripheral vision. A concept can be semiactive by fading from active status through lack of recent mention or it can be an essential but not yet activated element in an active frame an INFERRABLE CONCEPT concept; Prince 1981:236, Chafe 1996. Chafe also uses the term ACCESSIBLE for semiactive concepts, but others use this term in the sense of “identifiable.” Semiactive concepts seem to always be in a space that is current to the addressee, either the current discourse space or a base space. • An INACTIVE or NEW CONCEPT is one which is not present in a current space but which is to be set up as a result of the utterance or of some other act of ostensive communication; it might also be a nonessential part of an open space Chafe 1996:46. It is like a light that might or might not be installed but in any case is not turned on, or like something may or may not be there but which is completely outside of our vision. Activation status, like consciousness itself, is a matter of degree. Concepts that have been activated but are not further mentioned will fade from consciousness and lapse into semiactive status in around two to four utterances. For the speaker to maintain them in active status, he generally needs to refer to them within this interval. This requires only minimum coding for the language, or even less if certain conceptual conditions are met: paragraph topic §3.5.2, text-internal point of view §3.6.4. In general, as concepts become more active or have one of these other sustained conceptual statuses, their mention requires less coding weight; coding weight increases with effort expended in activation. In particular, expressions that have no sentence accent of any kind can generally be taken as designating active concepts Chafe 1994:75, Lambrecht 1994:95. 43 The converse does not hold: contrastive elements, for example, have at least a secondary sentence accent even when active: It’s all right with me if you go, but I’m not going. It is evidently possible for articles to signal activation status in attention management instead of identifiability in knowledge management. Levinsohn’s 2000, ch. 9 discussion of the definite article in Koiné Greek indicates that, as one dimension of its use and as a general rule with motivated exceptions, the article occurs with active and semiactive referents and is absent with inactive referents. 44 Thus, in James 1:3b–4a, ‘endurance’ is without the article when it is first introduced but takes the article once it is activated: ‘the testing of your faith produces Ø endurance. And let the endurance have its perfect 43 Cases of pragmatic accommodation §2.4.2, as in a story-initial sentence such as The old woman died discussed in Lambrecht 1994:197, can be taken as active in the sense that they are being treated as active and identifiable. 44 Thus the fact that nonidentifiable referents do normally not have the article is a result of their being inactive §2.4.2. Identifiable referents sometimes occur with the article and sometimes do not. result…’ NASB. Both the rule and its exceptions provide valuable evidence in regard to discourse topicality on different levels §3.5.2. Even though semiactivation relates to consciousness or attention and identifiability relates to knowledge, the two are often hard to distinguish and have large areas of conceptual overlap Chafe 1996, Mithun 1996. Semiactive concepts are generally identifiable and many identifiable concepts are also semiactive, but certain ones are not, such as probably, before it is mentioned the roof of the building you are currently in. As will be seen later on, both kinds of concepts are useful for topicality: all topics must be identifiable §3.1, paragraph topics remain active throughout their textual span, while remote topics are semiactive or active throughout their textual span §3.5.2. As noted earlier §2.3.1, the value for any relevant orientation dimension for a discourse unit appears always to be at least semiactive. A concept’s identifiability means that it is in some space that is current for the addressee, either in the space of the current discourse unit or in one of its base spaces such as world view §2.2.1. Much of world view takes the form of “frames,” which are schemas of conventional situations within the culture Croft and Cruse 2004:7ff.. When a frame such as RESTAURANT is active and instantiated, individual elements within the frame such as WAITER are identifiable. 45 Also, activating a referent appears to make other elements within its dominion identifiable, semiactivating many of them. 46 When one concept is used to make another one identifiable, it serves as a REFERENCE POINT in relation to the second concept Langacker 1998 and 2000, ch. 6. It is important to note that, in general, it is the activation of a reference point that makes concepts in its dominion identifiable. 47 Thus activation of one concept begets identifiability of related concepts, but activation does not in general beget activation nor does identifiability beget identifiability, except perhaps for very short “identifiability chains” in which there is an “essential and immediate kind of association” between the concepts Chafe 1996:46. If it were otherwise, activating just about anything would wind up activating everything else, and similarly with identifiability. To see limitations in identifiability chains, consider the decreasing identifiability of the underlined concepts in the following example, in which the slashes indicate alternative main clauses: As I was driving to work the other day, the car stopped working the battery came loose the positive cable came loose the bolt came apart. The car is identifiable from the “driving to work” frame, the battery would possibly also be identifiable, but not the positive cable or the bolt. To identify the bolt, it would be necessary to say something like the bolt on the positive cable of the car battery or the car battery positive cable bolt, that is, by grounding each successive referent by referring to some other identifiable referent: car → battery → positive cable → bolt. Activating an element makes elements in its dominion identifiable, and when those elements in turn are activated, elements in their respective dominions become identifiable as well, and so on recursively Langacker 1998. What kinds of concepts can be activated? Besides referents, it appears that predicators and propositions can be activated—more precisely, the conceptual events and states that lie behind predicators and propositions Chafe 1994:120. When active, all three can be referred to by minimal reference in English, pro-forms or, exceptionally, Ø: Referents: I know her; I know her but Ø don’t like her. Predicators: I know her and you do too; I like her, and you Ø her sister. Propositions: I know it; I know Ø. 45 Frames set up expectations as well as making referents identifiable. The following example, cited by Croft and Cruse 2004:14 is due to Fillmore: I had trouble with the car yesterday. The ashtray was dirty. Activating the car as a frame does make the ashtray identifiable, but the car trouble frame also makes this text noncoherent because it sets up limited expectations for things that make for “car trouble,” and dirty ashtrays are not included. 46 This is Fauconnier’s 1997:41 Access Principle, earlier called the called the Identification ID Principle 1985[1994]:3: “an expression that names or describes an element in one mental space can be used to access a counterpart in another mental space.” Similar terms are “inferrable” concepts, “evoking,” “bridging,” and “accessing”; see Brown and Yule 1983:79, Croft and Cruse 2004:13f., and papers in Fretheim and Gundel 1996. 47 Cf. Givón 1992:41: “A file can only be reached and opened by activating its label.” center of attention ⊃ active ⊃ familiar ⊃ uniquely identifiable ⊃ referential ⊃ type identifiable it that, this, this N that N the N, proper noun indef. this N a N For all three kinds of concepts, active status is implied by the absence of sentence accent, although, according to Lambrecht 1994:112, with predicators the correlation is not as strong as with referents, or possibly with propositions. A proposition is treated as pragmatically presupposed if the speaker assumes it is already stored in the addressee’s mental representation §2.3.3, hence can be cited and used without challenge, without being further asserted. Assertion is an instruction for the addressee to process and store the proposition Givón 1989:135. In other words, a presupposition is already stored in mental representation, while an asserted proposition is not Lambrecht 1994:52. The distinction between assertion and presupposition for propositions is a state of knowledge that corresponds exactly identifiability for referential enties: both presupposition and identifiability assume that the concept has been stored. Presupposition has been widely discussed in the literature. The activation state of propositions, however, has been relatively neglected in the literature. For example, it has only recently been recognized that in a so-called focus- presupposition sentence such as MARY kissed John, the presupposed component “X kissed John” needs to be active as well as presupposed Dryer 1996:476. In English, active propositions appear to be spoken without sentence accent, as in MARY kissed John, I KNOW it, I KNOW Ø, or in an utterance such as I KNOW she left you , where she left you has no sentence accent. If she left you were to have an accent, it could be an inactive proposition, in an exchange such as the following: A: Mary’s just away for a few days visiting her mother; B: Don’t try to fool me, I know she LEFT you with sentence stress on left. Presupposition, in this treatment as in Lambrecht 1994:52, has to do with the addressee’s knowledge of a proposition rather than her belief in or acceptance of it. Addressee’s belief seems to have no counterpart in the case of referents, and belief seems to be independent of activation. Consider the following conversation: A: What do you know about physics? B: I believe that E = mc 2 , just like you do. In this example, E = mc 2 is cited as an inactive proposition, with primary sentence accent, yet it is believed. In another conversation, this same proposition could be active but not believed: A: You don’t really think that E = mc 2 , do you? B: I don’t believe it any more than you do. So the fact that a certain proposition is active for the addressee is independent of her belief in that proposition, but either active status or belief appears to imply that the addressee knows about the proposition presupposition.

2.4.2 The Givenness hierarchy