2.1 What is open reference?

According to Siewierska (Reference Siewierska2004:1–2), person as a grammatical category covers the expression of the distinction between the speech act participants (that is, the speaker of an utterance and the addressee), and the party talked about that is neither the speaker nor the addressee. The category of person from a pragmatic and grammatical point of view thus prototypically involves the three-way distinction of speaker, hearer and third party (ibid.). However, languages typically have more than just three person markers, and in addition, these person markers can be used to refer not only to a specific individual or a group of individuals but to people in general or to a loosely specified collective (Siewierska Reference Siewierska2004:210).

It has been recognized in a wide variety of languages that person markers may also be used with no clear reference to a specific individual or group of individuals. The non-specific use of person markers can be found among the singular as well as non-singular personal forms. Within European languages, it is well known that the 2nd person singular forms are often used to create non-specific reference; this phenomenon has been studied in the Border Karelian dialects of Finnish (Uusitupa Reference Uusitupa2017), Danish (Jensen Reference Jensen2009, Nielsen et al. Reference Nielsen, Fosgerau and Jensen2009), Dutch (de Hoop & Tarenskeen Reference de Hoop and Tarenskeen2015), English (see e.g. Kitagawa & Lehrer Reference Kitagawa and Lehrer1990, O'Connor Reference O'Connor1994, Kamio Reference Kamio2001, Stirling & Manderson Reference Stirling and Manderson2011), French (Williams & van Compernolle Reference Williams and van Compernolle2009), German (Bredel Reference Bredel, Friedrich Graumann and Kallmeyer2002, Malamud Reference Malamud2012, Kluge Reference Kluge2016), Spanish (Posio Reference Posio2016), and [Finnish] Swedish (Fremer Reference Fremer, Kotsinas, Stenström and Drange2000). Furthermore, Siewierska (Reference Siewierska2004:212) mentions that the non-specific – or impersonal, as she calls it – use of the 2nd person singular is documented in Slavic languages, in Hungarian, Estonian, Komi, Turkish, and Abkhaz. In addition to European languages, the non-specific use of the 2nd person singular occurs in, for example, Godie, Gulf Arabic, Hindi, Kashmiri, Koromfe, Koyra Chiini, Kurdish, Mandarin (see also Biq Reference Biq1991), Marathi, Mauwake, Maybrat, Macushi, Modern Hebrew, Mundani, Nkore-Kiga and Tuvaluan. The non-specific use of the 1st person singular, on the other hand, has so far been much less studied (see, however, Helasvuo Reference Helasvuo2008 for Finnish, and Zobel Reference Zobel2016 for German).

From a typological point of view, the non-specific use of the 3rd person singular is considerably less frequent than that of the 2nd person singular (Siewierska Reference Siewierska2004:212). There are nevertheless non-specific 3rd person singular forms as well, such as the reflexive impersonals in, for instance, Romance and Slavic languages: the referents of these forms are necessarily human and they are often used in contexts in which the speaker is included or could be included (see e.g. Siewierska Reference Siewierska, de los Ángeles Gómez González, Lachlan Mackenzie and González Álvarez2008: 18–21).

Among non-singular forms, the non-specific use of the 3rd person non-singular is rather common across languages. Such use is found in many Indo-European language families, such as Germanic, Romance and Slavonic languages and, in addition, some other Indo-European languages such as Greek, Kashmiri and Persian, as well as in some Uralic languages both in the Finno-Ugric language family (in the Ugric, Permic and Mordvinic languages and in Mari language) and also among the Samoyedic languages in Nenets. Furthermore, the non-specific use of the 3rd person singular is documented in the Turkic languages, in the Dravidian languages, in some Niger-Congo languages, in some Trans-New Guinea languages and in some Austronesian languages (Siewierska Reference Siewierska2004:211; see also Siewierska Reference Siewierska, de los Ángeles Gómez González, Lachlan Mackenzie and González Álvarez2008). Interestingly, 2nd person non-singular forms tend not to be used for non-specific reference at all (Siewierska Reference Siewierska2004:211). Within European languages, the 1^st person non-singular forms are used for creating non-specific reference, but they do not appear to be as common outside Europe (ibid.).

In addition to the non-specific use of person markers, many languages also have certain personal constructions that are specialized in expressing deictically non-specific reference, such as the English one, French on, Spanish uno, or Germanic man constructions (see e.g. Altenberg Reference Altenberg2005 Ragnarsdottír & Strömqvist Reference Ragnarsdóttir and Strömqvist2005, Siewierska Reference Siewierska, de los Ángeles Gómez González, Lachlan Mackenzie and González Álvarez2008, Posio Reference Posio2016). Kibort (Reference Kibort2008) has specified different constructions in Polish that lack a canonical subject, and Leinonen (Reference Leinonen1985) in Finnish and Russian. In relation to modal verbs and necessity, Zinken & Ogiermann (Reference Zinken and Ogiermann2011) have studied the Polish impersonal modal declarative trzeba ‘need to’ construction that cannot be combined with a grammatical subject at all.

In previous studies, there has been variation in terminology when describing the phenomenon in question. Some studies speak about a ‘generalized’ (e.g. Stirling & Manderson Reference Stirling and Manderson2011 on English conversations) or ‘generic’ reference (e.g. Fremer Reference Fremer, Kotsinas, Stenström and Drange2000 on Finland's Swedish, Kamio Reference Kamio2001 on English, Jensen & Gregersen Reference Jensen and Gregersen2016 on Danish), while some have chosen to use the term ‘impersonal reference’ in their description of the phenomenon (e.g. Kitagawa & Lehrer Reference Kitagawa and Lehrer1990 on English, Siewierska Reference Siewierska, de los Ángeles Gómez González, Lachlan Mackenzie and González Álvarez2008,Footnote ² Malamud Reference Malamud2012 on German; for a discussion, see also Gast et al. Reference Gast, Deringer, Haas and Rudolf2015:149; De Cock & Kluge Reference De Cock and Kluge2016:352).

In this study, we use the term ‘open reference’ instead of generic, generalized, or impersonal reference. By using the term ‘open’ we would like to draw attention to the fact that even the so-called generic or impersonal expressions might find their referent(s) in the immediate speech context they appear in (see Laitinen Reference Laitinen2006, Helasvuo Reference Helasvuo2008). The use of open reference thus leaves space for the fact that expressions can be simultaneously non-specific and specific in their contexts of use, as Laitinen (Reference Laitinen2006:216) notes. Expressions that convey an open reference are non-specific in the sense that they are commonly used to present a generic or a generalized situation that anybody, at least in principle, can relate to. However, open expressions can also be rather specific in the way that they often identify a very particular experience that has happened to somebody, for example to the current speaker or to a certain third party who is not present in the speech situation, and in these cases, the context of use might allow a rather specific interpretation (see e.g. Laitinen Reference Laitinen2006:218–219, 224). With regard to referentially open expressions, it is thus the speech situation, the ongoing sequential action, and the responses that the participants’ produce in the situation that in the end define whose territory of experience is being addressed (see Heritage Reference Heritage, Stivers, Mondada and Steensig2011 on the territories of experience).

2.2 The zero-person and open 2nd person singular in the Finnish person marking system

The category of person in Finnish can be expressed in three coding systems: personal pronouns, verbal person markings, and possessive suffixes (Helasvuo & Laitinen Reference Helasvuo and Laitinen2006:173).Footnote ³ The person marking system is flexible in such way that it partially allows the so-called pro-drop: on the syntagmatic level, it is possible to leave out the 1st and 2nd person subject pronouns, and in such cases the marking of the verb – or that of the possessive suffix – conveys the personal reference (ibid.:174, 179), whereas the 3rd person pronouns can only be left out in specific contexts (ibid.:182–183). However, as Helasvuo & Laitinen (ibid.:179) note, in the spoken language the subject pronoun is most often present (see also Helasvuo & Kyröläinen Reference Helasvuo and Kyröläinen2016, Väänänen Reference Väänänen2016).

When placed in the paradigm of the Finnish person marking system (presented by Helasvuo & Laitinen Reference Helasvuo and Laitinen2006), the open 2nd person singular, from a grammatical point of view, falls into the same category as the deictically specific 2nd person singular, while the zero-person can be understood as a personal category of its own. This is due to the fact that the zero-person has a specific grammatical marking, as was mentioned in the introduction: in the zero-person construction, the predicate is in the 3rd person singular form and there is no overt grammatical subject (Laitinen Reference Laitinen2006:10). The modern examples of the use of the open 2nd person singular, on the other hand, seem to favor the presence of both the pronoun and the verbal person marking (Seppänen Reference Seppänen2000; Helasvuo & Laitinen Reference Helasvuo and Laitinen2006:201; Leino & Östman Reference Leino, Östman and Leino2008:39–40).

The zero-person has traditionally been considered a non-specific member of the Finnish personal system (see Hakulinen Reference Hakulinen, Verschueren and Bertuccelli Papi1987, Helasvuo & Laitinen Reference Helasvuo and Laitinen2006), whereas the status of the open 2nd person singular is not yet as conventionalized, at least in standard Finnish. However, as Uusitupa (Reference Uusitupa2017:37) notes, on the basis of the grammatical descriptions of Finnish written during the 19th and 20th century, the open use of the 2nd person singular seems to be rather common in the Eastern dialects of Finnish, especially in the southeastern region of Finland (see also Surakka Reference Surakka2011, Uusitupa Reference Uusitupa2011). The recent studies focusing on contemporary colloquial Finnish suggest that the open use of 2nd person singular is also becoming more common in spoken Finnish outside the eastern varieties of Finnish (see e.g. Laitinen Reference Laitinen2006, Suomalainen Reference Suomalainen2015, Suomalainen Reference Suomalainen2018).

As mentioned, the grammatical manifestation of the zero-person and open 2nd person singular is different, but the two constructions have similar semantical potential: they can both be used in a generic way, describing common or non-specific human experiences, and more specifically in their immediate context, so that they refer to the action, thoughts or experiences of a certain people or of the speech act participants (Helasvuo & Laitinen Reference Helasvuo and Laitinen2006:202; Laitinen Reference Laitinen2006:218–219, 229; Suomalainen Reference Suomalainen2018). It is, however, worth noticing that they take their referents distributively unlike, for instance, the unipersonal passive in Finnish whose implied referent is typically collective (Laitinen Reference Laitinen2006:218).Footnote ⁴ Example (3) illustrates the use of the zero-person (line 1) and the open 2nd person singular constructions (line 2).

1. (3)

   

Since the 2nd person singular can be indicated with a pronoun that can be case-inflected, the open 2nd person singular can, unlike the zero-person, be used more explicitly to indicate object and possessive forms in addition to the subject position. According to Laitinen (Reference Laitinen2006:213), the status of the zero-person as an argument in oblique cases is somewhat questionable, even though such cases have, in some studies, been analyzed as potential zero-persons (e.g. Vilkuna Reference Vilkuna1989:48–49, 194–195).Footnote ⁵

Because there cannot be a subject NP in the clauses with zero-person subject, the preverbal elements in zero-person clauses are more varied than in open 2nd person singular clauses, where the preverbal position is usually filled by the pronominal subject. Furthermore, it has been argued that Finnish could be considered as a topic-prominent language in the sense that, in an active sentence, the argument functioning as the topic need not to be the subject NP, but can also be for instance an adverbial or an object (Holmberg & Nikanne Reference Holmberg, Nikanne and Svenonius2002:78; for Balto-Finnic passives and impersonals and their relation to verb types and preverbal elements, see Hiietam & Manninen Reference Hiietam, Manninen, Heinat and Klingvall2005). Hakulinen & Karttunen (Reference Hakulinen, Karttunen, Corum, Cedric Smith-Stark and Weiser1973) have outlined the types of preverbal elements, such as adverbials and object NPs in zero-person clauses (or ‘missing person sentences’, as Hakulinen & Karttunen call them); the preverbal element can be, for instance, an expression of time, place, instrument, manner or a goal, and it forms the necessary or sufficient conditions for the process described in the clause (Vilkuna Reference Vilkuna1992:171–175; Laitinen Reference Laitinen2006), that is to say the referent of the preverbal element affects anyone that is in the particular situation.Footnote ⁶ In this respect, open 2nd person singular clauses differ from the zero-person clauses since the 2nd person singular pronoun (sinä or its variants), whenever present, tends to hold the theme position and thus function as a preverbal element (Suomalainen Reference Suomalainen2015:68–69). However, if the open 2nd person singular clause does not have a 2nd person pronoun subject, the preverbal element might be missing or be something other than a pronoun, for instance an adverbial or an object NP.

From the point of view of their contexts of use, the zero-person and the open 2nd person singular in a subject position have affinities: both are typical in hypothetical contexts as well as with modal verbs of necessity and possibility (Laitinen Reference Laitinen2006:212; Suomalainen Reference Suomalainen2015:66–67). Semantically, the person implied in zero-person constructions has a tendency to have the role of the beneficiary, the experiencer or the patient as the construction is often used to express changes of state (paleltua ‘to freeze’), emotions (iloita ‘to be happy’), perceptions (nähdä ‘to see’), experiences (viihtyä ‘to enjoy’), losses (menettää ‘to lose’), receptions (saada ‘to get’) and dynamic modality (päästä ‘get’). Agentive and stative verbs are possible mainly with a modal verb (as the infinitive complement of the modal verb), in a conditional frame such as if–then clausal compounds or in a generic complex sentence. (Laitinen Reference Laitinen2006:212–213). The open 2nd person singular, on the other hand, can be used more freely with agentive and stative verbs (Laitinen Reference Laitinen2006:219).

In conclusion, despite the similar potential for open reference, it has been argued that the use of the zero-person has certain semantico-grammatical restrictions that the open 2nd person singular lacks (Laitinen Reference Laitinen2006:219). Later in this article in Sections 4 and 5, we will reflect on these observations in light of our data and examine them more closely.

4.1 Grammatical factors

We analyzed the main predicate verbs of ØSG3 and OSG2 clauses in our data, mainly with reference to their semantics and argument structure. We followed the verb type classification of Pajunen (Reference Pajunen2001) but, in addition, we used transitivity and semantic proto-roles of proto-patient and proto-agent (see Dowty Reference Dowty1991) as classification criteriaFootnote ⁸ in order to better reflect the observations concerning the semantics of the verbs used in zero-person clauses (see e.g. Laitinen Reference Laitinen1995, Reference Laitinen2006). In our data, we consider three main groups of verbs: 1) concrete, 2) mental and 3) modal. Concrete verbs are further classified into verbs of (i) location, (ii) action, (iii) motion and (iv) event and change. Mental verbs consist of (i) psychological verbs, (ii) verbs of perception and (iii) speech act verbs. In Pajunen's (Reference Pajunen2001) classification, concrete and mental verbs represent primary verbs, whereas the secondary verbs contain the modal verbs and the aspectual verbs. In our study, only the modal verbs represent the secondary verbs. In the case of an aspectual verb, such as alkaa ‘to begin’, in the finite form, we have analyzed its infinitive complement as the main verb.

Marking the context as hypothetical or nonfactual is crucial when construing open reference, especially in zero-person clauses. In addition to the preverbal theme discussed in Section 2.2 above, other ways of creating a hypothetical or nonfactual context are, for instance, the conditional if–then frame (see Laitinen Reference Laitinen2006:212, 215; Helasvuo & Vilkuna Reference Helasvuo and Vilkuna2008:232) or the conditional mood. Examining the conditions which make the open reference of the subject possible, we took three factors into consideration: (i) the elements in the initial field, (ii) conditional conjunctions jos ‘if’ and kun ‘when’, and (iii) the conditional mood. The elements in the initial field are in this study divided into adverbials and object NPs.

As mentioned, the verb type classification partially depicts the semantic proto-roles of agent and patient, but in order to further investigate these roles we wanted to pay attention to the transitivity of the clause. The concept of transitivity in grammar and discourse is a complex one and its definitions vary (see e.g. Hopper & Thompson Reference Hopper and Thompson1980:251–252). In this study, we have adopted the view of Helasvuo & Kyröläinen (Reference Helasvuo and Kyröläinen2016) and encoded transitivity as the realization of the object complement (nominal or infinitival, not clausal) in our list of variables.

Some observations on tense and its effect on the interpretation of a zero-person clause have been made in earlier studies. It has been noted that with a past-tense form the zero-person clause is more often understood to be referentially specific (see e.g. Laitinen Reference Laitinen2006:212–213). In order to discover whether there is a difference in the distribution of tenses between our two personal constructions, we compared the present tense to the tenses that locate an action or an event prior to the moment of utterance.

Finally, we examined polarity in our data. Recent studies on conversational Finnish show how the 1st person singular subjects favor verbs of cognition (Helasvuo Reference Helasvuo2014:64), but there are also observations on how the most common cognitive verbs in Finnish conversational data are skewed with respect to person, number, and polarity (Laury & Helasvuo Reference Laury and Helasvuo2016:82). As Laury & Helasvuo (Reference Laury and Helasvuo2016:83) show, even the 2nd person singular negative forms seem to be used significantly less than those of the 1st person singular among cognitive verbs. We endeavored to discover if there are any such differences between the ØSG3 and OSG2 clauses in our data. All variables are summed in Table 1.

5.1 Adverbials and object NPs in ØSG3 and OSG2 clauses

In the statistical analysis, adverbials and object NPs in the initial field as well as the realization of the object NP on the whole proved to be statistically significant. The adverbials and object NPs in the initial field strongly favor the ØSG3 clauses, while the object NPs in general are more common among OSG2 clauses. This means that the ØSG3 clauses more often have an object NP, as in (4a) below, an adverbial, as in (4b), or both in the initial field.Footnote ¹⁰ This is not particularly surprising, as clauses with a zero-person subject in Finnish cannot have a subject NP, and therefore the initial field is left open for other elements (Laitinen Reference Laitinen2006:214–215).

1. (4)

   

In our data, object NPs in the initial field are rare among the OSG2 clauses (N = 9, of which six are relative pronouns that are obligatorily clause-initial). As mentioned in Section 4.2, in OSG2 clauses, the theme position is often occupied by the 2nd person pronoun sinä ‘you’, as in (5a) below. Open 2nd person singular subjects can also be expressed through the verbal marking only, and it would thus be possible in OSG2 clauses to have a preverbal object NP or an adverbial, or to leave the initial field open. However, if there is an object NP or an adverbial in the initial field, it can either be in the pre-field if there is also a subject pronoun, as in (5b), or in the theme position if there is no subject pronoun, see (5c).

1. (5)

   

As stated in Section 4, there are more object NPs among OSG2 clauses than among ØSG3 clauses, but these object NPs are typically placed in a post-verbal position, as is seen in (6).

1. (6)

   

It is noteworthy that the object NP distribution is largely in line with the differences among the mental verbs: 55.8% of OSG2 clauses with a mental verb have an object NP, while only 39.7% of ØSG3 clauses have an object NP.Footnote ¹¹ According to our interpretation, the reason for this might be that the open 2nd person singular construction, to a greater extent, is used in rather concrete discourse contexts; even when it is used with mental verbs, the processes described with these verbs need to be as concrete as possible. We will go deeper into this in Section 5.3, where we deal with the distribution of verb types among the ØSG3 and OSG2 clauses from a sequential point of view.

Generally speaking, the positioning of the clausal constituents in the ØSG3 and OSG2 clauses has to do with the information structure of the clause. As Vilkuna (Reference Vilkuna1992:9) notes, the word order, that is to say the constituent order at the clause level, is grammatically relatively free in Finnish, but discourse-conditioned. In other words, despite the existence of a grammatically unmarked default order, choosing a different order indicates something about the information structure and discourse functions of the clause (ibid.; Hakulinen et al. Reference Hakulinen, Vilkuna, Korhonen, Koivisto, Heinonen and Alho2004:§1366–1367). The information structure correlates with the structure of the object NP in the ØSG3 clauses of our data: in the initial field, up to 64% of object NPs are demonstrative pronouns, while in positions other than the initial field, approximately 62% are lexical nouns. On the conversational and pragmatic level this may reflect the use of the ØSG3 clauses when referring to something already introduced into the discourse, as in (7a), or spatially connecting the utterance to the speech situation, is shown in (7b).

1. (7)

   

5.2 Conditional mood in ØSG3 and OSG2 clauses

In this section, we provide a qualitative observation of the occurrences of the conditional mood in regard to ØSG3 and OSG2 clauses, as well as the interactional contexts in which the clauses with conditional mood occur. Although in the scope of this study it is not possible to exhaustively categorize the conditional clauses of our data (N = 240), some remarks can be made on why this particular mood is so common among ØSG3 clauses and not as widely used in OSG2 clauses. We argue that the typical uses of the conditional mood in the Ø3SG clauses in our data have to do with (joint) planning and proposal making, seen in examples (8)–(9) below, and indicating wishes, seen in (10). Moreover, with the conditional mood the speaker may express her stance on the topic under discussion, as in (11)–(12). All these uses are discussed later in in this subsection.

The conditional mood in Finnish differs from that of many other languages (Kauppinen Reference Kauppinen1998:156). In Finnish, the conditional mood has been seen to indicate nonfactuality (see e.g. Kangasniemi Reference Kangasniemi1992:242–243; Kauppinen Reference Kauppinen1998:156–160). The conditional mood also has more pragmatic, secondary functions, for example in conveying politeness, doubt or mitigation (Yli-Vakkuri Reference Yli-Vakkuri1986:191–201). In some subordinate clause positions the conditional mood has been said to correspond the subjunctive mood (see Kauppinen Reference Kauppinen1998:164).

Diachronically, non-actual planning – that is nonfactual per se – has been considered as one of the prime functions of the conditional mood. This is due to the alleged combination of frequentative derivational suffix and the past-tense suffix in the conditional marker -isi- (Lehtinen Reference Lehtinen1983:485–501). According to Hakulinen (Reference Hakulinen, Verschueren and Bertuccelli Papi1987:149), the conditional mood in zero-person clauses conventionally implicates the speaker her/himself as the implied agent/experiencer in clauses which express either the speaker's plans or intentions, or enable an optative-like interpretation in which the speaker indicates a wish of some sort. Optative-like conditional clauses are typically verb-initial but may also include a conjunction (e.g. jos ‘if’) in the beginning of the clause. In such turns, the speaker often seeks for an approving response from other participants – even if the speaker refers to her/himself (Kauppinen Reference Kauppinen1998:187–190).

Example (8) demonstrates how the zero-person forms of verbs of action and movement in the conditional mood are used in joint planning as Mikko proposes a trip to a sauna to a group that consists of his wife and another married couple (see also Couper-Kuhlen & Etelämäki Reference Couper-Kuhlen and Etelämäki2015:12).

1. (8)

   

Besides the concrete verbs, also the modal verbs, especially the verb voida ‘can’, occur typically in contexts of planning along with a zero-person form. Consider example (9), in which Iina is planning a trip to Turku with her husband, who is not present in the conversation.

1. (9)

   

Whereas the concrete verbs and the modal verbs were used in (joint) planning in examples (8) and (9), it seems that the conditional mood marks the speaker as the implied agent/experiencer particularly in connection with the mental verbs, in example (10).

1. (10)

   

In contexts like this, the speaker typically indicates her wishes or intentions. In example (10) the zero-person also alternates with the singular 1st person.

Both personal and impersonal pronouns provide a way to express stance: personal pronouns often express a personal or otherwise involved stance whereas impersonal pronouns fit better with less personal stance (van Hell et al. Reference van Hell, Verhoeven, Tak and van Oosterhout2005). Expressing stance, and especially potentially intersubjectively shared stance, seems to be rather common among ØSG3 and OSG2 clauses as well. Consider example (11), where the speaker provides an evaluation of a TV program about off-road motorcycle racing.

1. (11)

   

The nonfactuality indicated by the conditional mood is thus used to convey the speaker's stance as he argues that one needs a custom-built motorcycle in the ongoing race instead of an ordinary motocross motorcycle.

It is noteworthy that in all of the OSG2 clauses that have a predicate verb in the conditional mood the nonfactual interpretation is supported by contextual cues such as a conditional frame or a rather crystallized 2nd person singular imperative form of the verb ajatella ‘to think’, as seen in (12). In (12), the speaker again expresses her stance on the topic under discussion.

1. (12)

   

We believe that the key to the differences in the frequency of the conditional mood of ØSG3 and OSG2 clauses lies in the very essence of the conditional mood: If we consider it as a marker for nonfactuality, it does not alone provide sufficient conditions to avoid a misunderstanding of the referent in OSG2 clauses. Instead, the speaker has to mark the nonfactuality in some other grammatical or contextual way, and therefore the conditional mood would in many cases be redundant.

5.3 Verb types in ØSG3 and OSG2 clauses

As stated in Section 4, the perception verbs and verbs of event and change slightly favor the ØSG3 clauses, whereas the speech act verbs as well as the verbs expressing action, motion and location occur more often in the OSG2 clauses. Our qualitative analysis shows that this kind of distribution of verb types can be explained by two factors: Firstly, the semantico-grammatical restrictions zero-person has (described in Section 2.2 above) make certain interactional contexts more suitable and some impossible for the use of this construction. Secondly, the interactional tasks the ØSG3 and OSG2 clauses have in their conversational contexts are of importance, and the type of action of the ongoing sequence has an impact on which of the construction types is being used.

The tendency that perception verbs and verbs of event and change are slightly more common in ØSG3 clauses can be approached through the observations presented in earlier studies; it has been stated that the zero-person often has the proto-patient role of implication of affectedness, so that the person implied in the ØSG3 clause is often in the role of the beneficiary, experiencer or patient (Laitinen Reference Laitinen2006:213). This also seems to hold true for the ØSG3 clauses in our data, in which the implicit subject in ØSG3 clauses observes, feels and is affected by the circumstances.

The use of the open 2nd person does not have similar semantico-grammatical restrictions to the zero-person. Nevertheless, it seems that the OSG2 clauses have a tendency to have a certain role in the sequence in which they occur. In our data, the 2nd person singular with an open reference is typical in contexts where the speaker demonstrates something with a concrete example. The extract in (13) is from a conversation between three young adults, who are talking about what a curriculum vitae is and what information it is supposed to include. Just before the extract, Kaisa has told the others that she is not quite sure what a CV is since she has never written one. However, in lines 1–2, Kaisa tries to define the contents of a CV, and Masa, who has just written his first CV, joins in.

1. (13)

   

In (13), the OSG2 clause with an action verb kirjoittaa ‘to write’ occurs in Kaisa's turn in line 2 when she starts to explain what she thinks one – or more precisely ‘you’ – can write in the CV. In lines 3–4, Masa, in overlapping talk, completes Kaisa's definition and explains what information a CV is supposed to include. In his turn, Masa also uses the open 2nd person singular, but as an adverbial expressing the possessor in a possessive clause (line 4). In terms of interaction, the OSG2s in (13) occur in a context in which both participants are engaged in trying to give a definition on the topic of the talk, in this case the content requirements of a CV.

Furthermore, in our data, the ØSG3 and OSG2 clauses co-occur relatively often. This means that the speakers seem to use the open 2nd person singular in the same sequences where they also use the zero-person. In sequences where ØSG3 and OSG2 clauses co-occur, the distribution of tasks is the following: the ØSG3 clauses are used to give a general characterization or to introduce a topic or a change of perspective, while the OSG2 clauses are used to illustrate (the previously described) state of affairs or to exemplify a claim (see Nielsen et al. Reference Nielsen, Fosgerau and Jensen2009:126–129 on the distribution of tasks of the two Danish generic pronouns). This distribution of tasks is also visible in the statistical analysis; verbs of event and change are more frequent in the ØSG3 clauses, whereas verbs describing dynamic action or being in a location favor OSG2 clauses.

Example (14) below demonstrates how the distribution of tasks of the OSG2 and ØSG3 is reflected in the types of verbs used with each construction type. The extract comes from a conversation between four young men who are having a game night. The topic in (14) is music festivals, and the participants are discussing whether people are allowed to bring their own drinks to the Tuska festival. Before the following extract, Heikki has told the others how, due to changes in the law, people are not allowed to bring their own drinks into the festival area anymore. As a response, Tuomas, starting from line 1, seeks more clarification about this claim.

1. (14)

   

In (14), the ØSG3 clauses occur in lines 1–2, 8–10 and 11–12, in both Heikki's and Tuomas’ turns. All these turns describe the general states of affairs. The first ØSG3 clause in Tuomas’ turn in lines 1–2 deals with the way the implied person(s) act in the outlined circumstances. It is a question of hypothetical situation (note the jos ‘if’ frame), and the context makes the intentional interpretation, in which Tuomas might be planning his future actions, possible with regard to the use of the zero-person. In the example above, we can also see how the zero-person is typical with modal verbs: in both Tuomas’ turn in lines 8–10 and Heikki's response in lines 11–12, the generic agent implied with ØSG3 is under the modal condition of permission, which is expressed with the modal verb saada ‘to be permitted/allowed to’.

Consequently, in (14), the ØSG3 clauses are used to frame the topic situation. The OSG2 clauses, on the other hand, are used in those parts of the sequence where Heikki moves on to argue why it is best to buy the three-day ticket to the Tuska festival. In lines 3–7, in the turn that contains the OSG2 clauses of this example, Heikki provides a reasoning for his argument; he gives a concrete example of how one can act if they have a three-day ticket. The subordinating conjunction, the complementizer että (line 4), marks what follows it as a paraphrase of the prior talk (see Koivisto, Laury & Seppänen Reference Koivisto, Laury, Seppänen, Laury and Suzuki2011:71) and thus highlights the demonstrating and illustrative task of the turn.Footnote ¹² The verbs used in his turn describe agentive action (jättää ‘to leave’, käydä ‘to go’, hakea ‘to get, to pick up’, mennä ‘to go’, juoda ‘to drink’). In addition, the open 2nd person singular forms in Heikki's turn are used in an affective context that involves evaluation, and they function as a part of an assessment sequence (see e.g. Goodwin & Goodwin Reference Goodwin, Goodwin, Duranti and Goodwin1992). In this context, Heikki's assessment sequence could also be interpreted as functioning as a piece of advice to Tuomas who is a bit hesitant to buy the three-day ticket to the festival; in this case, the OSG2 forms in (14) can be understood as carrying a somewhat hearer-directed meaning at the same time as they are generalized descriptions of a certain way of acting.

Interestingly, in (14), Tuomas in his response turn (lines 8–10) does not respond to the affective content of Heikki's turn at all, but instead asks, with the help of the zero-person, whether one can still bring water into the festival area. With his turn and the zero-person construction in it, Tuomas thus brings the conversation back to a general level, which Heikki then continues in his following response in lines 11–12, where he also moves on to using the zero-person. In this context, the change in person form from OSG2 to ØSG3 not only signals a change in the ongoing action, but also a change in perspective and footing (see Goffman Reference Goffman1981).