Minimal reference: The use of pronouns in Finnish and Estonian discourse

Front matter

| Studia Fennica. - ‘ Linguistica

THE FINNISH LITERATURE SOCIETY (SKS) was founded in 1831 and has, from the very beginning, engaged in publishing operations. It nowadays publishes literature in the fields of ethnology and folkloristics, linguistics, literary research and cultural history.

The first volume of the Studia Fennica series appeared in 1933. Since 1992, the series has been divided into three thematic subseries: Ethnologica, Folkloristica and Linguistica. Two additional subseries were formed in 2002, Historica and Litteraria. The subseries Anthropologica was formed in 2007.

Contents

MARJA ETELAMAKI Context and referent in interaction Referential and indexical dimensions of the Finnish demonstrative PYONOUNS ... csse

EEVA-LEENA SEPPÄNEN Pronouns, gaze and reference The Finnish demonstrative pronoun tama as a device for modifying participation frameworks in conversation . . .

Ritva LAURY First and only

Appendix

Morphological glosses ABL ablative

ACC accusative

ADE adessive

ALL allative

CLT clitic

COMP comparative COND conditional ELA elative

ESS essive

GEN genitive

ILL illative

IMP imperative

INE inessive

INF infinitive

NEG negation

PAST past tense

PCP participle

PL plural

PSS passive

PTC particle

PTV partitive

PX possessive suffix Q question

SUP superlative

TRA translative COM comitative

ISG first person singular (likewise for 2"* and 3")

first person plural

Context and referent in interaction Referential and indexical dimensions of the Finnish demonstrative pronouns

In this paper, my aim is to describe the referential and indexical features of the Finnish demonstrative pronouns tä(m)ä, toi (tuo) and se, with emphasis on the pronoun tä(m)ä.! I examine everyday talk as activity, where artefacts are constituted and talked about along with all sorts of other things being done. I assume that conversational referents are best understood as interactional entities (see Etelämäki 1998); furthermore, they only come to being in interaction when referred to. In addition, they are identified in relation to the on-going activity, regardless of whether they are concrete, physical objects, or linguistically created ones (i.e. second or third order entities, Lyons 1977: 443-445)? As Heath & Hindmarsh (2000) write, the objects are “inseparable from the environment in which they are located and the specific courses of action in which they figure”. Hence, also referential expressions become meaningful only in and through the interaction. Furthermore, the profoundly interactional character of referents-in-interaction is embodied in the grammar of reference forms, particularly in the semantics of demonstrative pronouns.

The acts of reference are profoundly reflexive: the objects of reference are identified in relation to the on going activity, and simultaneously the activity is reflected and constituted through the references to the objects. Therefore, referential expressions themselves take part in indexing the activity, and in that way create their own indexical contexts. When considering referential expressions, the reflexive nature of reference invites us to look at two directions, namely the referential and the indexical. Demonstrative pronouns are referential indexicals, which means that their basic function is to individuate objects of reference in terms of their relation to the indexical ground of reference (Hanks 1990: 36-43;

1 The paper is based on my forthcoming doctoral dissertation Tarkoite ja toiminta. Tutkimus suomen pronominista tä(m)ä.

2 This is not an ontological claim about the existence of entities per se, but a claim about the existence of mutually shared referents in interaction.

1992). The indexical ground is that part of the interactional context that functions as the sociocentric origo for the reference. Since the context is dynamic, the demonstratives function in two ways: while referring to an entity they also organize the indexical ground, and set a relation between the referent and the indexical ground (Heritage 1996 [1984] 236—237, see also Hanks 1992: 53).

I approach linguistic phenomena through ethnomethodological conversation analysis, where conversation is examined as organized activities accomplished in and through turns-at-talk. In conversational data, the unfolding interaction appears as the primary context for the referential acts. The data used for this paper is naturally occurring face-to-face, and telephone conversations. In the next section, I will describe the Finnish demonstrative system. I will then proceed with the analyses of three conversational sequences to exemplify the use of the pronouns, and to empirically argue for the referential and indexical features that I propose in this paper. Finally, I will briefly discuss the idea of indexical context including conversational activities.

The Finnish demonstratives

There are three demonstrative pronouns in Finnish, namely tä(m)ä, toi (tuo), and se. Besides the three demonstrative pronouns, there are also pronominal adjectives and adverbs, which are based on the same roots.

Table 1. The Finnish demonstratives.?

The Finnish demonstrative pronouns

Pronouns:

tää (tämä) ( this’) nää (nämä) (”these”)

toi (tuo) ( that’) noi (nuo) ( those’)

se Cite the’) ne Cthey’/ the’) A sample of other Finnish demonstratives

Adjectives: Adverbs (Locatives): Adverbs/Particles:

tä-mmönen (tä-llainen) tää-llä näin to-mmonen (tuo-llainen) tuo-lla noin se-mmonen (se-llainen) sie-llà niin

The translations above are only approximate ones, since the Finnish reference system is quite different from the English one. All the Finnish demonstrative pronouns may be used either independently or adnominally. There are no unambiguous articles in Finnish, but adnominally used pro-

3 The first form is commonly used in southern Finnish vernacular. The standard Finnish forms are in parentheses.

nouns se (Laury 1997: 250—263) and even tä(m)ä (Juvonen 2000: 196) have been noted to possess features of a definite article.

In my description, I will follow Hanks’ (1990, 1992) model of demonstrative reference, first applied to Finnish by Laury (1997). He proposes that the meanings of referential indexicals are composed of three features, namely characterizing, relational, and indexical. The characterizing features designate properties of referents, the relational ones convey the relation of the referent to the ground, as well as ways of access (tactual, visual, discourse) to the referent, and the indexical features organize the indexical ground against which the referent is to be identified (Hanks 1990: 66, 1992: 51-53). According to Laury (1997), the characterizing features of the Finnish demonstratives designate the referent to be either “the one", i.e. a single object, or “the region". The relational features express that the referent is either included in or excluded from the context. The indexical features organize the context to either speaker (+ addressee) or addressee centric (1997: 58-62). My proposal differs somewhat from the ones introduced above, as can be seen from the table below. I have only specified the referential (i.e. characterising) and indexical features in the table, since they are in focus in this paper:

Table 2. The Finnish demonstrative roots.

Roots: Referential features: Indexical features: tä- open asymmetric

t(u)o- open symmetric

se- closed symmetric

For the sake of uniformity, I use the term referential instead of characterizing, when referring to the features that correspond to referential dimension. I suggest them to be “open” (for further definitions) and "closed", for the following reasons. As mentioned above, the demonstrative class in Finnish includes pronouns, adverbs, and even adjectives. They are all based on the three demonstrative roots fä-, to-/tuo-, and se-. The pronouns refer to a single entity, while the adverbs might refer to a region (täällä, tuolla, siellä), manner (näin, noin, niin, täten, siten), or time (tällöin, tuolloin, silloin), and the adjectives to guality* (tällanen, sellanen, tollanen). In these forms, such referential properties as regional, manner, or temporal are designated by affixes that are mostly frozen forms of

4 The pronominal adjectives are often used as determinants for potentially referential NPs, in which case they can be compared to other NP determinants. However, as NP determinants they emphasize the category that the NP expresses rather than the specificity of the referent (see Vilkuna 1992: 132-133; Juvonen, in this volume), which is due to their adjectival character.

original case markers. Hence, it is not the demonstrative root that makes these distinctions. *

In addition, the root itself seems to convey some referential meaning. It designates the referent properties that have to do with the issue of definiteness. In the next section, I will propose that definiteness, when understood as *knownness”, is not a static feature of a referent, nor is it a dichotomy (see also Du Bois 1980; Chafe 1994: 93-100; Laury 1997: 34—51). The question is not only whether the referent is identifiable, but also whether it is adequately identified for the on-going activity. The pronouns fä(m)ä and toi express that paying attention to, describing, or determining the referent is somehow relevant for the on-going activity, and therefore the referent is open for further characterizations. That way they also function as pointing. The pronoun se marks the referent to be sufficiently defined for the on-going purposes, and therefore "closed" for further identification.

The indexical grounding of referents simultaneously makes reference to, and articulates with, the context in which the reference is performed (Hanks 1991: 254—255). Thus, the indexical component of reference organizes the indexical ground of reference. I propose that the Finnish demonstratives organize it as either “symmetric” or “asymmetric”, according to the mutuality of the participants’ understanding of the ongoing activity, rather than proposing the reference to be either speaker or recipient centric. This is because in my view the indexical ground is profoundly social (see below; also Hanks 1990: 36—43), and based on the on-going interaction. In the next two sections I hope to exemplify the points I have made so far, by analysing three extracts from actual conversational data. I will first discuss the referential features, and then the indexical ones.

Referential features

In this section, the main emphasis will be on the characterizing features of the pronoun tä(m)ä, but I will discuss the other two demonstratives, too. I have argued that the pronouns tä(m)jä and toi convey that their referents are still open for discussion, whereas the pronoun se marks its referent as known enough for the on-going purposes. Since td(m)d and toi point at their referents as “observables”, their use is parallel to the

5 Inaddition to referring to inanimate referents, the Finnish demonstratives might be used for referring to persons, and even to the participants of a conversation (Seppánen 1998; Kaiser, in this volume). Among other things, they differ from the Estonian demonstratives in this respect (see Pajusalu, in this volume). Since the demonstratives might be used to referring to human and inanimate entities, the referential features do not designate such features as “animate” or “human” either. However, in person references the pronouns do propose particular participant statuses for the referent (Seppänen 1998). In addition, a notion of participant roles is related to the notions of humanity or animacy.

act of pointing. I will exemplify my proposal in the analysis of the following extract (1).

The extract is from a conversation with two participants, Airi and Sisko. They are both artists, and they are at Airi’s studio discussing her paintings that are present in the room. The extract is from the beginning of the situation. I will focus on Airi’s turns, and show how the demonstrative pronouns function, first, in directing the recipient to look at the painting, and secondly, in organizing the unfolding activity.

(1) (Airi ja Sisko)

Airi:

Sisko:

Airi:

(3.2)

Sisko:

Airi:

Sisko:

Airi:

ensin puhutaan tosta?.hh first let’s talk about that one?

mm:. mhm.

(0.5)

h::$(h)y(hy k(h)yl seon aika kaamee.hh$.hhniin h::$(h)i(hi it(h) is guite awful indeed.hh$ totah (2.4) se nyt jatkaa sitä lootustee?maa?? so:h (2.4) it now continues the lotus theme?

mitä lootusteemaa what lotus theme sitä mitä mä tein (.) siihe viime näyttelyyn the one that I did (.) to the previous exhibition sen pallon (.) pyóreen jos oli se valkonen that ball the round one that had that white loo[tus keskellà? (0.5) (0.2) mut]ta: (.) se on lo[tus in the middle? (0.5)|(0.2) but] (.) it has

[NODDS REPEATEDLY]

(.) jostain syystäh (.) ruvennu mua kiinnostamaan (.) for some reason (.) begun to interest me tää (0.6) pelkän tumman ja vaaleen (0.4) #niin#ku this (0.6) pure dark and light (0.4) #like# vaihtelu? variating?

15 (.)

16 Sisko: mm. mm.

17 (0.4)

20 Airi: ja silleen että (.) toi ei niinku (.)tavallaan ois and so that (.) that would not like (.)in a way be

21 mitään ainetta toi kukka. (4.8) et et sillon se ei any like material that flower. (4.8) that that then

22 oo niinkun mikään esine. >tai se ei oo< kukan it is not like any object. >or it is not< a picture

23 kuva oikeestaan vaan sitä on vaan of a flower really but it has only been

24 sitä kukkaa käytetty (1.2) niinku tohon that flower used (1.2) like for the purposes of that 25 (1.4) sommitteluunh (1.4) compositionh 26 (2.0) ((Sisko walks towards the painting and stops (3.6) ))

28 Airi: [mitä raati sanoo [what does the jury say

In the extract, Airi introduces the first painting to be discussed about. It begins with Airi pointing at the painting (line 1), and ends with an inguiry after the recipient’s opinion (line 27). The introduction can be divided roughly in three parts. In the first part, Airi points to the painting and proposes it to be the first one to be talked about, and Sisko aligns with the proposal (lines 1-2). Then, Airi gives subsidiary information about

6 Apparently, Sisko’s movements (line 23), and faint laugh (line 24) prompt the actual guestion. However, both the movement and the voice can be taken as expressing her orientation to the situation, i.e. the introduction being taken to its end, and thus to a transition relevance place.

the painting (lines 4-10). Finally, she offers a framework for mutual examining of the painting (lines 10-24).” * The turn in line 1 initiates the talk about the paintings:

1 Airi: ensin puhu-taan to-sta?.hh first talk-3PL that-ELA first let's talk about that one?

2 Sisko: mm:.

PCT mhm.

The formulation of the turn serves this purpose. The adverbial ensin (‘first’) is in the theme, by which I mean the position that in a clause internal syntactic structure precedes the verb, and is most often occupied by a subject in active transitive clauses (Vilkuna 1995). It is followed by the verb puhutaan (‘talk’) that denotes to the activity itself. The verb is in passive form, which in spoken Finnish can be used as a first person plural imperative. Hence, the the verb form makes the utterance a proposal. The painting is pointed to by the pronoun tosta (an elative form of the pronoun toi). It organizes the indexical ground as symmetric, and simultaneously points at the referent as open for further discussion. By symmetric ground I mean that both the participants have equal access to the referent, due to their mutual understanding of the on-going activity. The activity itself — "talking about the surrounding paintings" — is shared in a sense that it is what the participants are oriented to doing in this particular situation. This includes that they share the understanding of what the role of the referent is in the activity.

7 The dashed line (above lines 9-10 in the transcript) marks gaze. The object of gaze is written in the beginning of the line. For the sake of readibility, I have only transcribed gaze where it is particularly relevant for this analysis.

8 In these data, there are sequences of talk that are organized around the co-developing of a topical line. However, it seems that constructing and assessing a referent in these sequences is an activity that the participants are oriented to; hence, the talk is not organized around topical development but around the activity (see Sacks 1992 [1968], April 17). The demonstratives place the referent primarily within the activity, and indicate how the utterance is related to the unfolding activity simultaneously as they function in organizing the activity itself. In addition, indexes seem to cumulate, and hence it would be an overstatement to claim that it is the demonstrative alone that marks the turn/utterance to do such and such a thing. E.g. in extract 1 line 5 there is 1) the utterance's position (after the pointing to a referent), 2) the demonstrative pronoun se, and 3) the particle nyt, that all together produce a particular interpretation of what the utterance is doing, and what its sequential implications are. There might even be grounds to ask, whether the demonstrative is actually doing anything else than expressing a relation of a referent to a particular kind of context, where the demonstrative would be chosen due to the context. However, there is a reflexive relation between an utterance and its context, i.e. the context is set through the utterance itself, and one cannot tear these apart or put them in a causal or temporal order.

There is no speaker change after the pointing. By a minimal response mm. (line 2), Sisko agrees to talk about the painting that Elvi points to, but withdraws from taking the next turn. Ochs Keenan & Schieffelin (1983 [1976]) have suggested that a discourse topic is a proposition rather than a single entity. In a way, their observation is evidenced in my data, where a pointing of a referent usually co-occurs with a proposition concerning the referent. The proposition provides for a mutual point of view to the referent: a framework, which provides grounds for understanding further talk. It is the pointing together with the proposition that composes a topic offer. The turn in line 1 only consists of a pointing. Therefore, I suggest that although there is a transition relevance place after line 1, there is also an implication that Airi will continue talking. I understand Sisko's minimal response that passes the turn as expressing her orientation to that implication. In lines 4—5 Airi takes a turn again:

4 Airi: —h::$(h)y(h)y k(h)yl se on aika kaamee.hh$.hhniin PTC it be+3sc quite awfull PTC h::$(h)i(h)i it(h) is quite awful indeed.hh$

5 totah (2.4) se nyt jatka-a si-tà lootustee°ma-a°? PTC it now continue-3sc that-Prv lotus theme-Prv so:h (2.4) it now continues the lotus theme?

In line 4, she first produces a negative assessment about the painting. The assessment is signalled as a parenthetical comment in various ways. To begin with, the painting is being referred to by a pronoun se (see the analysis of se below). In addition, the assessment is framed with laughter that signals the utterance as non-serious.? Finally, after the assessment the speaker rushes on to an in breath and produces a particle chain niin tota, which indexes return to the main line, to an activity projected by the earlier talk (Heinonen 2002: 110-113). In doing that it projects more talk by Airi. By producing the particle chain right after the in breath, the speaker passes a possible transition place following a grammatically and pragmatically complete utterance. That way she expresses that the previous utterance — although an assessment — was not initiating an assessment sequence (for assessments, see Pomerantz 1984: 57-101, Goodwin & Goodwin 1987, 1992).

After a pause that follows the particle chain, she begins to tell about the painting. However, by the pronoun se and the particle nyt, the new utterance (se nyt jatkaa sitä lootusteemaa) is indexed as delivering subsidiary information. The pronoun se refers to the painting first as a subject in the theme, and then as a part of an object-NP sitd being the partitive

9 Whatis being done with a laughter-framed, self-deprecating comment in this position would be a subject of larger analysis than what I offer here. However, the subject is not within the scope of this paper.

10 Heinonen (2002) discusses cases where niin tota is in a second pair part, indexing return to an activity projected by the first pair part.

case form of se (sitä lootusteemaa).'' It designates its referent to be an entity that is adequately identified for the purposes at hand, and needs no further discussion (see Laury 1997: 58-128). Thus it expresses that the utterance does not offer the painting to be talked about within the framework of “lotus theme”.

The particle nyt supports the interpretation of the information as subsidiary: nyt can mark that a clause is a part of the ongoing activity, but is not taking the main line of the talk forward (Hakulinen 1998). Together with the pronoun se, the particle marks the utterance as one that offers background information. Since the utterance follows a (mere) pointing of a referent, but the indexical expressions se and nyt mark it as expressing subsidiary information, there still is an implication that the speaker will continue talking. This is an implication that the participants orient to.

Contradictory to my claim, there is a pause (line 6), which ends only when Sisko takes the next turn (mitä lootusteemaa) in line 7:

6 (3.2)

7 Sisko: mitä lootusteema-a what lotus theme-Prv what lotus theme

However, her turn does not take the talk forward, but functions as a next turn repair initiator. The NP sitä lootus teemaa is used as a recognitional (Himmelman 1996: 230-239; for se, see Laury 1997: 102-103, 118), which makes either a confirmation or a repair initiation by the recipient a relevant next action (cp. also *try-marking”, Sacks & Schegloff 1979). Himmelman (1996) describes recognitional use as such where “the referent is to be identified via specific, shared knowledge rather than through situational clues or reference to preceding segments of the ongoing discourse". He mentions three features that are attached to the recognitional use, all of which fit to the above example: 1) the identification of the referent may cause problems for the recipient, which is oriented to by the speaker, 2) the referent is often of only peripheral importance, and 3) additional information which makes the referent more accessible is often added in relative clauses or other modifiers of similar complexity. (Himmelmann 1996: 230.)

It must be noted that recognitional use can't be taken as an inherent property of the pronoun se, since other reference forms can be used as recognitionals, as well. Furthermore, se also has non-recognitional uses, for example tracking (the first se in line 5). However, Himmelmann's observations are useful in this analysis: the new referent (lotus theme)

11 Though the clause is a transitive clause, its function is rather to express a description of the subject-NP than an event where the subject-NP would be an agentive participant, and the object-NP a goal. The subject-NP functions as an object of description, and the verb together with the object-NP compose the description. Hence it is functionally close to predicate nominal clauses.

is introduced as identifiable, and it is offered a subsidiary role in further talk. In addition, the pause leaves space for the recipient to confirm that she recognises the referent in question. In this extract, the recipient initiates a repair sequence (line 7).

The trouble source for the repair initiator (line 7) is the “lotus theme”. As soon as the initiator of the repair (Sisko) has confirmed (by nodding, line 11) that she has identified the referent, the speaker introduces a new aspect to the painting, and the repair sequence becomes a side sequence. The continuation (beginning by mutta ‘but’ in line 10) can be interpreted either as turn-internal or turn-initial, depending on whether Sisko’s nod counts as a turn or not. On the one hand, it is a confirming response, and acting as a sequence closing third to the repair sequence. On the other hand, the response being only a nod — and not a verbal response — it can be seen as an orientation to the incompleteness of the on-going turn. The continuation (line 10, 12—14) introduces another angle to the painting:

10 loo[tus keskellä? (0.5) (0.2) mut]ta: (.) se on but it be-3sc lo[tus in the middle? (0.5)|(0.2) but] (.) it has 11 Sisko: [NODDS REPEATEDLY ]

12 Airi: —(.) jostain syy-stéh (.) ruven-nu mu-a kiinnosta-ma-an for some reason-ELA begin-PCP 1sc-PTV interest-3INF-INE (.) for some reason (.) begun to interest me

13 tää (0.6) pelkän tumma-n ja vaalee (0.4) #niin#ku this pure-GEN dark-GEN and light-GEN PTC this (0.6) pure dark and light (0.4) #like#

14 vaihtelu? variation variating?

Sorjonen (1989: 174—176) has pointed out that mutta (‘but’) often initiates turns that do not continue "talking on topic", but continue "talking topically". Typically, they are used for introducing new sub-topics to the conversation (ibid). Talking topically refers to talk that merely involves attention to topical coherence, whereas talking on topic consists of “blocks of talk about ‘a topic?” (Sacks 1992 [1968], April 24). In extract 1, the utterance (in lines 10, 12-14) continues topical talk about the painting, but does not continue on the topic of the previous utterance (“lotus theme").

In addition, the utterance is syntactically constructed to accomplish a transition in the course of the talk. There are two co-referential NPs in the utterance (lines 10, 12-14). The former (se) is an argument of the verb, and the latter (tää pelkän tumman ja vaaleen niinku vaihtelu) is structurally a right-dislocation.? In my data, constructions with a right-dislocated NP determined by the pronoun tä(m)ä are used for re-defining the topic." If there is a pronoun se as the first NP (in the argument position), the utterance begins a new sub-sequence in an activity that is going on already. In extract 1, the participants are already talking about a particular painting. The pronoun se indexes the continuation of the overall activity, and points to its referent as sufficiently defined for the momentary purposes: a topical referent for the on-going talk. The right-dislocated NP specifies the reference by expressing a new category that the referent belongs to.

The pronoun tää in the dislocated NP designates the referent to be open for further discussion, which means that it offers the referent for further characterizations. The categorization functions as a framework that continues in the following characterizations of the painting. The pronoun also organizes the indexical ground as asymmetric. That way it indexes that there is a change in the activity, as well: the utterance shifts the talk from delivering subsidiary information about a topical referent to offering a framework (“variation of light and dark") for discussing on topic.

The turn in lines 20-26 continues discussing the painting in the “dark and light"-framework:

20 Airi: ja silleen että (.) toi ei niinku (.)tavallaan o-is and so+that that that NEG--3sG PRT in-a-way be-COND+3sG and so that (.) that would not like (.)in a way be

21 mitään ainet-ta toi kukka. (4.8) et et sillon se ei any material-Prv that flower that that then it NEG+3sG any like material that flower. (4.8) that that then it

22 oo niinkun mikään esine. >tai se ei oo< kuka-n be like any object or it NEG+3sG be flower-GEN is not like any object. >or it is not< a picture of

23 kuva oikeestaan vaan si-tà on vaan picture really but it-PTv be+3sc merely a flower really but it has only been

24 si-tä kukka-a — käyte-tty (1.2) niinku to-hon that-PTV flower-PTV use-PSS+PCP PTC that-iLL that flower used (1.2) like for the purposes of that

12 The terms “co-referential” and “dislocation” are only used to clarify the syntactic construction in question, and to refer to particular elements in it. However, I find literal understandings of these terms highly questionable.

13 The data consist of video recorded face-to-face conversations (approximately 45 minute conversation between the two artists, and half an hour television talk show) and audio recorded everyday telephone conversations (approximately one and half hours).

The reference to the flower is first determined by the pronoun foi, in the utterance that explicates the abstractness of the motif by negating the “materiality” of the flower. In the continuation, the flower is referred to with the pronoun se, which marks it as sufficiently characterized. Instead, in the NP that refers to the composition (tohon sommitteluun) in the last utterance, is determined with the pronoun toi (tohon in the illative case). The pronoun foi designates that the referent is yet open for further characterizations. And it is the composition — consisting of the variation of dark and light — that the speaker offers to be commented on.

To sum up, the painting is first referred to by the pronoun toi, where both the participants know what the activity will be, and the referent is still to be discussed. When the speaker is talking about an aspect of the painting that is not offered for further discussion, she uses the pronoun se to refer to the painting. Se is also used adnominally in an NP that introduces a subsidiary aspect of the painting. If an NP introduces an aspect that is offered for the following talk, it is determined by the pronoun tad. The pronouns tää and toi maintain the referent under observation, and the pronoun se merely tracks a referent in the conversation. In addition, the pronoun tää makes identifying — in the meaning of ‘knowing about’ — the referent particularly salient in the act that is the turn is accomplishing. Therefore, I propose that the referential feature of tä(m)ä and toi is "open", and se is “closed”. As has been implied already in this section, the pronoun tää does not only introduce a new topic framework, and offer it for further talk. It also indexes re-organizing of the activity. This is due to its indexical feature, which I will discuss in the next section.

Before turning to the indexical features, I will briefly compare my view of the pronoun se to the views of the other authors in this book, and discuss the notion of definiteness in the light of interaction. Tiainen- Duvallon (2002, in this volume) claims that the pronoun se is the most neutral reference form: there are no additional meanings or functions attached to its use besides reference.'* Furthermore, she brings out that the use of the pronoun se is not constrained by a lexical mention or the presence of the referent, but the identification of the referent is guided by the host construction of the pronoun, and its position in a larger linguistic context (in this volume). Although for Tiainen-Duvallon, it is the syntactic construction that provides for the identifiability of the referent, her findings fit well with my views. In conversation, actions are accomplished in and through linguistic constructions. Hence, recognizing the activity and recognizing the construction are just two sides of the coin: using a

14 This comes close to what Schegloff (1996) writes about practices for referring to persons in talk-in-interaction. He asks: “How do speakers do reference to persons so as to accomplish, on the one hand, that nothing but referring is being done, and/or on the other hand that some thing else in addition to referring is being done by the talk practice which has been employed? Relatedly, how is talk analyzed by recipients so as to find that ‘simple’ reference to someone has been done, or that referring has carried with it other practices and outcomes as well?" Following Schegloff's line of thought, one could say that with the pronoun se, nothing but referring is being done.