Conjunction

Conjunctions form one of the metatextual classes. As such, they operate within the themerheme structure. Just like particles, conjunctions add a commentary to the rhematic component. However, given that they bind to other units on both sides and operate on complex T-R structures, their commentary – unlike that of particles – affects the relationship between the theme and the rheme parts of the T-R structure, connected to the former by the common main theme.

Many properties of conjunctions result from their metatextual nature, such as 1. their lack of involvement in the clause structure (conjunctions are not required by any element of the clause and do not determine the way in which open positions will be filled); 2. the fact that they do not take any non-contrastive sentence stress or 3. cannot be negated. Their participation in a multi-component structure of knowledge makes them more stable than particles in terms of their position. These two properties, i.e. the involvement in complex T-R structures and the related property to bind with two, rather than one – as in the case of particles – element, combined with the positional stability of conjunctions in contrast to the mobility of particles, are currently the criteria used to distinguish between these two, most researched to date, metatextual classes.

1. Conjunctions in Polish research. From markers of reference to metatextual commentaries

Traditionally, conjunctions have been defined as invariable non-independent words with a connective function. As such, they are contrasted with prepositions and relative proforms. The differentiation between conjunctions and prepositions relied on the following four criteria: 1. the type of phrases being connected (prepositions combine heterogeneous segments of a clause, while conjunctions combine homogeneous clause segments or entire clauses (Szober 1923, 1957; Laskowski 1984; Saloni, Świdziński 1998); 2. the manner of connection between such elements (prepositions determine subordinate relationships, while conjunctions both the coordinate and the subordinate ones, see Szober 1957; Jodłowski 1971, 2003; Klemensiewicz 1963; Wajszczuk 2011); 3. government (prepositions govern the case, see Szober 1957; Jodłowski 1971; Saloni 1974; Grochowski 1984) and 4. syntactic independence (preposition – together with the noun it governs – fills a syntactic position, see Klemensiewicz 1937, 1963; Wajszczuk 2011). The discussion on the distinction between the two classes has continued, in a way, along two parallel paths. On the one hand, it revolves around the second criterion, i.e. the relationship established by conjunctions, and its combination with the first criterion – the type of phrases being connected (Szober, Jodłowski, Laskowski). On the other hand, it focuses on the way in which units representing both classes participate in the sentence syntax (criterion 4.) (Klemensiewicz, Laskowski, Wajszczuk). Within both currents, the distinction between prepositions and conjunctions is the weakest at the level of a single sentence (combining parts of a clause), but the inclusion of the syntactic criterion 4. allows for distinguishing between the two classes while recognizing the similarity of the relations that they introduce (cf. also Grochowski 1984; Frankowska 1995; Gębka-Wolak, Moroz, Wiśniewski 2006).

The connecting nature of relative pro-forms (in Polish: relatory) (also referred to as relative connectors (in Polish: konektory względne) Laskowski 1984) and their functional comparability to conjunctions was first observed by Zenon Klemensiewicz (cf. connection markers). Nevertheless, this similarity was soon challenged from the syntactic perspective by Tadeusz Milewski (1965) (conjunctions connote two positions without determining how they are filled, while relative pro-forms refer to a main clause segment). The grammars that followed also specified that relative pro-forms, unlike conjunctions, may take the position of a component of a clause (cf. Bobrowski 1995; Wróbel 1996; Grochowski 1984; Laskowski 1984, 1998; Ulitzka 2010; Bednarczuk 2016). See also the analysis of conjunctions which contain a pro-form-like element (Urbańczyk 1939; Cyra 2000; Sadecka-Makaruk 2004; Sosnowska 2011). For more on the relationship between relative pro-forms and conjunctions in Polish linguistics (Grochowski 2020).

A change of the syntactic perspective related to the shift from searching for dependencies between the elements of a clause towards recognizing conjunctions as a commentary on the T-R structure (Wajszczuk 1997, 2011) allowed scholars to better articulate their intuitions on the independence of conjunctions from the structures that they connect (e.g. Klemensiewicz 1963). At the same time, this modification of the research perspective encouraged comparisons between conjunctions and other metatextual units, including particles, (cf. e.g.

Milewska 2006a; Danielewiczowa 2009).

It should be emphasized that traditionally Polish linguists have not challenged the fact that conjunctions have meaning in themselves (Bogusławski 1966, 1975, 1985; Wierzbicka 1969, 1972). Despite being recognized as elements with a connective function, the idea that their function is merely structural has not taken root in Polish academic circles.

Studies on specific conjunctions focus on their meaning, typically explored in connection with their syntactic properties (Wajszczuk 1997; Cyra 2000), with special emphasis on order (Grochowski 1977b, 2003; Sadecka-Makaruk 2004; Dróżdż-Łuszczyk 2006; Milewska 2006b), and on their origin (Pisarkowa 1984; Grzegorczykowa 2014; Kleszczowa, Termińska 2015). On top of that, drawing on Klemensiewicz’s concept of reference markers

(1949), researchers have analysed the impact of conjunctions on textual coherence (Górny 1960; Klemensiewicz 1966; Wierzbicka 1966; Pisarkowa 1984; Topolińska 1995; Bednarczuk 2016), including the coherence of spoken texts (Pisarkowa 1975; Ożóg 1990; Majkowski 2002). Sometimes scholars focus on the function of conjunctions in texts of a specific type, e.g. legal texts (Wierzbicka-Piotrowska 2016).

2. Subdivision of conjunctions

In the Polish tradition, the most established subdivision distinguishes between coordinating

(paratactic) and non-coordinating/subordinating (hypotactic) conjunctions (e.g. Szober 1923, 1957; Gaert ner 1938; Milewski 1965; Karolak 1993); more on the criteria for the classification of conjunctions into paratactic and hypotactic in European research in (Bednarczuk 1982/83). When it comes to coordinating conjunctions, we may distinguish between cumulative (conjunction), disjunctive (alternative), inclusive (equivalence and specification), adversative (opposition), and illative (implication) conjunctions, while the subordinating conjunctions may be further divided into: conditional (implication, counterfactuality) conjunctions, conjunctions of concession, cause-effect (causal), and sometimes also temporal conjunctions or the conjunctions of purpose. As Jadwiga Wajszczuk (1997; cf. also e.g. Bednarczuk 1967, 1982/83; Górska

2003) correctly noted, this classification is problematic for at least three different reasons. First of all, it is based on heterogeneous criteria, as it refers to the relationship between facts (e.g. cause-effect or illative conjunctions), between phrases (e.g. inclusive conjunctions) or, finally, to logical relationships (e.g. disjunctive conjunctions; for a discussion of the issues inherent to the application of logical categories to the description of conjunctions, (cf. e.g. Wojtasiewicz 1972; Magner 2005, 2015, 2016ab, 2017, 2018, 2019). The outcome is a non-homogeneous class consisting of highly diversified units. Secondly, as a result, the classification does not allow for a precise definition of the set of conjunctions. And thirdly, by defining subclasses on the basis of heterogeneous criteria, we risk ending up with the fictitious polysemy of the elements – one unit may be allocated to more than one class and thus may be erroneously considered as having more than one meaning.

A rarer subdivision of conjunctions is the one based on internal syntax. From this perspective, conjunctions are classified as either continuous (albo, mimo że) or non-continuous

(correlated conjunctions such as dopóki…, dopóty… (cf. Grochowski 1977ab, 1984; Kallas 1993; Sadecka-Makaruk 2004).

Wajszczuk (1997, 2011) proposed a syntactic classification of conjunctions taking account of their semantic properties. In her approach, the core group of conjunctions is made of conjunctions proper (in Polish: spójniki właściwe), subdivided into two groups a) conjunctions in multi-element orders (serial conjunctions) such as i, oraz, tudzież, a, lub, albo, czy, bądź, ni, ani and b) conjunctions in two-element orders, referred to by the scholar as central conjunctions, e.g. czyli, toteż, to,bo, gdyż, albowiem, lecz, ale. These units fill a fixed position between conjuncts. Unlike conjunctions proper, conjunctions ‘improper’ (in Polish: niewłaściwe) (skoro,

jak, jeśli, jeżeli, jeśliby, jeżeliby, gdyby, byle, byleby, chyba że, ponieważ, jako że, aczkolwiek, jakkolwiek, choć, chociaż) may, together with one of the conjuncts, fill any position with respect to the other conjunct. Because of this feature, in her later proposal (Wajszczuk 2011) she considered them as binding to one unit only and thus left them out of the conjunctions class.

A vast group of the abovementioned researchers have extended the class of subordinating conjunctions to also include complementizers such as że, żeby (e.g. Jodłowski 1951/71 considered them – along with gdy, jeśli, choć – as ad-clausal subordinating conjunctions, cf. Malejka 2003; Ćwiklińska 2007; Wójcicka 2017). However, their distinct nature within the class of conjunctions caught scholars’ attention very early on. For example, Tadeusz Milewski (1965) specified that they connote the main clause as a result of informative incompleteness. Likewise, Maciej Grochowski (1984) emphasized that they fill positions implied by other expressions (cf. Szpakowicz 1983; Karolak 1993; Wajszczuk 2011). Olgierd Wojtasie-

wicz (1972) explicitly pointed out to their substantial distinctiveness arguing that – unlike conjunctions – they do not contribute any specific meaning, but rather perform a structural function. According to Jadwiga Wajszczuk (1997), they form a class of complementizers referred to by the scholar as ‘includers’ (włączniki) – units characterised by the “extreme lack” of syntactic independence, implied by the predicate of the main clause and introducing the subordinate clause as the argument required by the main clause.

However, the grammatical classification of units such as zaś, natomiast, jednak, więc, zatem, bowiem, mianowicie remains open to discussion. Some of them (e.g. więc) were traditionally classified as conjunctions (illative ones in this case). In more recent approaches, they are considered as an intermediate class between particles and conjunctions (linking particles – Wajszczuk 1997). What they have in common with the former is their positional mobility, while they share with their latter their connective function. Researchers that classify them as particles rely on their function of referring to the previously communicated content (e.g. Grochowski 1996), while scholars who consider them as conjunctions explain their capacity to enter deep into a clause by the fact that – in the perception of the speakers – they maintain their (genetic) relationship to the primary modulating function (e.g. Milewska 2006b). None of these classifications is, however, satisfactory (e.g. Milewska 2007; Kisiel 2015). Assuming the criteria related to the role of conjunctions in STR and their positional properties,

units from this class cannot be regarded as conjunctions. Wajszczuk (2011) proposed considering them as a separate class but views them as most closely related to particles. This class (of ad-utterance operators) would contain one-place metatextual units, having positional freedom and requiring left-sided context based on their meaning. The scholar proposes to extend this class by including also some units traditionally considered as subordinating con-

junctions, such as skoro, jeżeli, jeśli, gdyby, chociaż, ponieważ.