On the functional relationship between the Czech reciprocity markers navzájem/vzájemně and spolu

1 Introduction

This paper is concerned with the relationship between the Czech adverbs navzájem/vzájemně ‘each other, mutually’ and spolu ‘with each other’, which indicate that the part of speech on which they syntactically depend, typically a verb used as a predicate, has a reciprocal meaning. As I argue in Veselý (2025), the two markers form a semantic opposition: navzájem/vzájemně indicates binary conjunctive reciprocity, while spolu indicates single-event reciprocity (König and Gast 2008: 24); both terms will be clarified below. Under certain semantic and pragmatic circumstances, however, they can become variants, i.e., it is possible to use both adverbs without a significant difference in meaning. This is not to say that their meanings are exactly the same. In my analysis, which is presented below, I have focused on those utterances found in the Czech National Corpus in which a reciprocal is combined with an adverbial marker that, at least on a general level, does not correspond to its semantic type. Hereafter I will call this (not corresponding) marker non-basic, whereas the corresponding marker will be called basic. In the research I focused especially on those cases where, according to the analyzed data, this non-basic use is much less common than the basic one, and although in such contexts the two markers become variants (it is always possible to substitute the non-basic marker by the basic one), subtle semantic differences between them can be observed. The aim of this paper is to describe these differences and to address the question of why this asymmetric variation occurs only with certain reciprocals used in certain contexts (and how it relates to the observed differences in meaning).

When studying the variation of units at any level of language, it is crucial to determine how common the alternatives are relative to each other, otherwise their functional relationship could not be described in a complex way. As I argue in Veselý (2025), when studying difficult cases of variation (i.e., variation of meaningful units in particular), a quantitative analysis of corpus data should be integrated with a qualitative analysis, rather than the latter being carried out after the former and independently of it. This can be done in a two-step manner: 1. as an analysis of a large amount of data, where the qualitative part must be undertaken to ensure that the frequencies have been correctly determined, i.e., that issues such as form-meaning asymmetries (polysemy and homonymy) and false syntactic pairs have been resolved and only true alternatives have been counted in the statistics, 2. as a detailed analysis of a limited number of utterances, which aims to ascertain the functional relationship between the variants with respect to their frequencies (obtained via the first step). The variation I am dealing with here was subjected to the first step of the process in Veselý (2025). The core of this text was a general picture of the variation between navzájem/vzájemně and spolu and methodological aspects of its research. This study, on the other hand, is an implementation of the second step, so the qualitative component of the analysis dominates over the quantitative one. Here, I will refer to the results of the general analysis presented in Veselý (2025) only to the extent necessary.

The corpus research was focused solely on the situation in the Czech language. I am not aware of other languages than Czech – except for Slovak, where the situation is very similar – that signal the opposition between single-event and binary conjunctive reciprocity by lexical (adverbial) means. While binary conjunctive reciprocity is often signaled by an adverb (e.g., mutually in English, einander, gegenseitig in German, etc.), single-event reciprocity is not signaled in this way, and it is often not manifested at all, i.e., not even by morphological or clitical means (e.g. they kissed, Die Eltern sprachen über alles Mögliche (German) ‘The parents talked about all sorts of things’, etc.). The Czech adverb spolu is unique in the sense that it is polysemous between a collaborative and a reciprocal sense (for example, the English adverb together does not display such a behavior). In this context it is worth noting the situation in Polish: while the adverb razem ‘together’ with a collaborative sense cannot also signal single-event reciprocity, the prepositional phrase z(e) sobą ‘with each other’ seems to be functionally similar to the Czech adverb spolu ‘with each other’, cf. Wiemer (1999: 308): “These observations are wholly compatible with the almost complementary distribution of jeden drugiego ‘one the other’ and nawzajem, wzajemnie ‘mutually’, on the one hand, and ze sobą, on the other: the former are used with canonical reciprocals, but excluded with verbs denoting naturally reciprocal events, whereas ze sobą may be used only with natural reciprocals and verbs related to them closely […].”^[1] It might be fruitful to undertake a crosslinguistic corpus-based research focused on a comparison of how the semantic opposition of single-event and binary conjunctive reciprocity is indicated in Czech, Slovak and Polish. Obviously, such research would require different methods than those used in this paper. Here I will only make a brief remark on the situation in Czech and Polish (another remark can be found in Section 3): the functions of the Polish z(e) sobą are distributed among at least three expressions in Czech: se sebou (bojoval sám se sebou / walczący z (samym) sobą ‘he fought with himself’), s sebou (vzali si s sebou svačinu / przynieśli ze sobą przekąskę ‘they took a snack with them’) and spolu (povídali si spolu / rozmawiając ze sobą ‘they were talking to each other’). The Czech phrase se sebou does not conventionally signal single-event reciprocity, but this usage is not entirely excluded, especially in co-occurrence with navzájem/vzájemně: stýkají se pouze se sebou navzájem ‘they only socialize with each other’.

The paper is organized as follows: Section 2 deals with the general concept of variation and how it is understood in this paper. Section 3 defines the category of reciprocity and outlines the classification of reciprocal expressions according to various criteria. Section 4 deals with the means of expressing reciprocity in Czech, and Section 5 focuses on their variation. In Section 6 the variation of the adverbial markers navzájem/vzájemně and spolu is analyzed in detail, especially the cases when a given marker is combined with a reciprocal of an “inappropriate” semantic type. Section 7 discusses the question of the semantic relationship between the two units.

2 Variation as a functional relationship of language units

Sankoff (1988) distinguishes three basic approaches to syntactic variation: 1) the introspective-generative approach studies variation as part of the language system; 2) the experimental-evaluative approach is concerned with cognitive processes related to the choice from the alternatives, using the method of psycholinguistic experiment; 3) the descriptive-interpretive approach analyzes syntactic variation based on language use, with data being mostly taken from linguistic corpora. The present study clearly adheres to the third paradigm of approaches.

According to the rigorous concept of variation, variants should have exactly the same meaning in a truth-functional sense, cf. Weiner and Labov (1983: 31–32): “[variants should be] truth-conditionally equivalent and used on the whole to refer to the same state of affairs”. Of course, this requirement is unproblematic when the variants are not meaningful units, i.e., in the case of variation between phonetic and graphic non-sign units. However, when it comes to variants at the morphological, lexical, and syntactic levels, the strict definition becomes difficult to maintain. In many conceptions, semantic differences are not only allowed but even assumed to be a necessary feature of any instance of linguistic variation, for example: “The hypothesis can be put forward that each grammatical form has a unique meaning not exactly matching that of any other one, not even that of its purported ‘variants’ or ‘alternatives’” (Aijón Oliva 2013: 574). In the context of construction grammar, much attention has been paid to variation of syntactic units, cf. Goldberg (1995: 67): “If two constructions are syntactically distinct, they must be semantically or pragmatically distinct”. It has been pointed out that because of these functional differences, each use of a given alternative is preceded by a meaningful choice^[2] made by the speaker in a given communicative situation. The concept of deliberate choice between nearly (but not completely) synonymous competing units, determined by a number of intralinguistic and extralinguistic variables, has been widely elaborated in the disciplines investigating language use (pragmatics, stylistics, and discourse analysis). On the development of research on grammatical variation towards this comprehensive view, see, for example, Aijón Oliva and Serrano (2012).

Some authors (e.g. Biber et al. 2021) treat variation as a relation between units of the language system, while others (e.g. Dorgeloh and Wanner 2010) use this term in reference to actual language use and with respect to genre-specific communicative settings. From a methodological point of view, I believe that one approach cannot be separated from the other: only through analyses of language use, we can “credibly” ascertain what is and what is not part of the language system, and what the relations between the “parts” are. However, it is possible to distinguish between a systemic variation, in which the alternation relation is part of the language system, and an occasional variation in individual texts.^[3] The latter type can be expected to manifest itself in significant differences between the frequencies of each variant, in contrast to the former type. Arguably, each type is associated with diverse perceptions of the functional differences by the speakers: the use of an occasional, infrequent variant may be perceived as a symptomatic substitution. However, a single alternation relation can sometimes be identified as systemic and sometimes as merely textual, depending on the context. For example, in Czech when the reciprocal hádat se ‘argue’ denotes an action with more than two reciprocants,^[4] the construction hádat se spolu ‘argue with each other’ commonly alternates with the construction hádat se navzájem ‘argue with each other’; this is therefore a systemic alternation. However, when the same reciprocal denotes an action with only two reciprocants, the construction hádat se spolu ‘argue with each other’ is only rarely replaced by the construction hádat se navzájem ‘argue with each other’; this is therefore an occasional alternation. I will discuss this issue in more detail below, especially in Section 7.

3 The semantic category of reciprocity and its expression in Czech

According to Haspelmath (2007: 2088), a mutual situation can be defined as a situation with two or more participants (A, B, …) in which for at least two of the participants A and B, the relation between A and B is the same as the relation between B and A. For example, the sentence John and Mary loved each other denotes a reciprocal situation, as indicated by the reciprocal anaphor each other, because not only John loves Mary, but also Mary loves John. Reciprocity/mutuality is related to symmetry. The predicate to love each other can be said to be symmetrical, because it consists of two (in the given example) relations that are principally the same: someone loves someone else (although each “someone” is assigned a different semantic role). However, the relation between the two notions (i.e., reciprocity and symmetry) is a matter of authorial conception. König and Kokutani (2006) define symmetry as a semantic property and reciprocity as a syntactic property.^[5] For example, the verb argue denotes an action that may not be perfectly symmetrical in meaning when the participants are represented in different valency positions (a argued with b), but it becomes perfectly symmetrical when used in a reciprocal construction, where the number of valency positions is reduced (a and b argued). However, constructions such as a argued with b are often recognized as a specific type of reciprocal, cf. the term “discontinuous reciprocal” coined by Dimitriadis (2008). Wiemer and Grzybowska (2015), on the other hand, distinguish between reciprocal and symmetrical predicates, see below in detail. In this text, I use the term “reciprocity” in the same way as Haspelmath uses the term “mutuality” (see above), i.e. as a semantic concept, and I assume that reciprocity is based on symmetry (which, however, can have different forms).

In Czech, reciprocal situations are most typically denoted by reciprocals containing the reflexive morphemes se, si; these expressions either form a lexical unit, as in example (1), or they are grammatical units^[6] semantically related to a non-reciprocal expression (in example (2), milovat se is a grammatical reciprocal based on the verb milovat):^[7]

These reciprocals are sometimes accompanied by adverbs indicating reciprocity, most typically navzájem/vzájemně ‘each other, mutually’ and spolu ‘with each other’ (however, the latter can also mean ‘together’, as in jeli tam spolu ‘they went there together’):

There can be various motives for using these adverbs; one is to eliminate ambiguity of the reflexive morphemes se, si; note that sentence (2) can also mean ‘John and Mary loved themselves’. The situation described by (2) can also be expressed by a grammatical construction containing the expression jeden druhý ‘each other’; in this case the marker navzájem/vzájemně cannot be attached to the reciprocal (however, exceptions to this rule can be attested in the corpus data):

König and Gast (2008: 24) distinguish between binary conjunctive and single-event reciprocity. In the case of binary conjunctive reciprocity, “a symmetric relation R is instantiated twice between two participants a and b (Masha and Vanya noticed each other)”, while in the case of single-event reciprocity, “two instantiations of a relation R are conceived of as a single event (Masha and Vanya kissed)”. The latter type corresponds to the class of predicates that Dimitriadis (2008: 376) calls “irreducably symmetric”; Kemmer (1993) calls them “natural reciprocals”. In Haspelmath (2007: 2106), the two notions are covered by the terms “uniplex mutual events” and “multiplex mutual events”: the sentence Ram and Dolores told each other a secret represents multiple sub-events (two telling events with two secrets involved), whereas the sentence Pedro and Aisha quarreled (with each other) describes only a single (i.e., uniplex) event. Leaving aside minor differences in the understanding of these concepts by individual authors, binary conjunctive / multiplex reciprocals can presumably be equated with grammatical (sometimes called canonical) reciprocals, and single-event / uniplex reciprocals with lexical reciprocals. These are natural correspondences: a lexical reciprocal lacks a non-reciprocal base verb, reflecting the fact that sub-events of a given reciprocal are not easily distinguishable. In this paper, I will use the notions single-event × binary conjunctive reciprocity / reciprocals, mainly for the following reasons: i) These are semantic concepts, and my analysis is mostly a semantic one (e.g., I work with the concept of neutralization of a semantic opposition, see Section 7); the terms grammatical × lexical reciprocals rather refer to different types of linguistic units. ii) I assume that the meanings of single-event and binary conjunctive reciprocity can sometimes be evoked by the reciprocal adverbs and other contextual clues, even though the reciprocal alone might suggest the opposite. And in some cases, it may not be easy to determine whether the meaning is only created in context or is associated with a reciprocal of a particular type. Having chosen the pair of terms above, this dilemma does not need to be resolved, i.e., the reciprocal does not need to be determined as a grammatical or lexical one.

It can be a matter of debate whether binary conjunctive reciprocals – and also all single-event reciprocals – are truly symmetric. According to Wiemer and Grzybowska (2015: 219–220), sentences such as Koni-e się kopi-ą ‘The horses kick each other’ contain a reciprocal predicate, not a symmetrical one: “In a reciprocal relation, there are two referents (individuals or sets thereof) each of which represents two different semantic functions, in other words: which are each aligned to two functions with complementary status in argument structure. It is important to realize that this involves two conditions: [a] X and Y stand in an identical semantic relation to each other, but [b] each of them is ascribed to two different arguments (roles). Symmetrical predicates fulfil the first, but not the second condition.” In sentences like Ich ręc-e się dotyka-ł-y ‘Their hands touched (each other)’ the predicate is symmetrical, since both conditions are met: “In these instances, X and Y (i.e. two different referents) can only be ascribed to one argument (i.e. semantic function).” Finally, single-event reciprocals like całować się ‘kiss (each other)’ represent intermediate cases, i.e., may show a transition between reciprocal and symmetrical predicates (Wiemer and Grzybowska 2015: 220–221). In my view, fulfillment of condition [a] alone is sufficient for qualifying a given predicate as symmetric: When two horses kick each other, the first horse relates to the second horse in an agent-to-patient way, and the second horse relates to the first horse in the same way. When two lovers kiss, they stand in a co-agent relationship with each other. And when two hands touch, each hand stands in a relation of spatial co-occurrence with the other hand. Both “parts” of the situation thus either can or cannot be assigned two different participant roles. Qualifying all these instances as symmetrical predicates does not mean that there are no significant differences between them. However, the broader concept of symmetry which I am endorsing here emphasizes the common features of both groups and reflects the unsharp boundary between them. The proximity of the two types also seems to be suggested by Kemmer (1993: 97, 117–118): whereas for prototypical (i.e., binary conjunctive) reciprocals, the initiator (i.e., agent) and endpoint (i.e., patient) roles are represented in their semantic scheme, for natural (i.e., single-event) reciprocals, there is no separate representation for these roles, which is consistent with Kemmer’s assumption of “less participant distinguishability between initiating and endpoint aspects [emphasis mine] of the participants”. Thus, even in co-agent cases such as całować się ‘kiss’, the action is “oriented” from one participant to the other and vice versa.

In Veselý (2025), I presented an analysis of data extracted from the SYN v8 corpus (Křen et al. 2019) to support the hypothesis that in Czech, both types of reciprocals are primarily marked with a different adverbial marker: binary conjunctive reciprocals are primarily combined with the adverb navzájem/vzájemně ‘each other, mutually’, while single-event reciprocals are primarily combined with the adverb spolu ‘with each other’. See examples (3) and (4); note that potkat se (spolu) ‘meet’ is a single-event reciprocal, and milovat se (navzájem/vzájemně) ‘love each other’ is a binary conjunctive reciprocal. In principle, my analysis of a) 20 most frequent collocations made up of a reciprocal containing the reflexive element se and the adverb navzájem/vzájemně, b) 20 most frequent collocations made up of a reciprocal containing the reflexive element se and the adverb spolu^[9] supported this hypothesis,^[10] i.e., examples (3) and (4) represent the typical, most frequent combinations. The result of the analysis will be specified in Section 6.

A crucial question is how it is possible to determine whether a reciprocal is a single-event or a binary conjunctive one. Several criteria are given below. It is not always (i.e., for each reciprocal) possible to apply all of them, and their application is not always entirely conclusive. The validity of the given criteria is not compromised by these obstacles, because i) individual reciprocals can show individual behavior, ii) the opposition between the semantic categories is not strictly contradictory, i.e., there is, arguably, a continuum of cases that cannot be clearly assigned to either category.^[11]

1. As observed by Siloni (2012), when a binary conjunctive reciprocal is determined by a multiplicative numeral such as pětkrát ‘five times’, ambiguity regarding the number of events arises; this is not the case when a single-event reciprocal is determined in the same way. For example: Jan a Petr se navzájem pětkrát navštívili ‘John and Peter visited each other five times’ = there were 5 visits / there were 10 visits × Jan a Petr se pětkrát potkali ‘John and Peter met five times’ = there were 5 meetings.

2. As mentioned above, single-event reciprocals are principally lexical reciprocals, and binary conjunctive reciprocals are grammatical reciprocals. Hence, the absence of a base verb indicates single-event reciprocity (assuming that the reflexive form is a reciprocal). However, it may not be trivial to determine whether the corresponding non-reflexive form is a base verb, i.e., whether the reciprocal has the same meaning when the reciprocal meaning is “subtracted”. For example, the reciprocal znát se is either a grammatical one, with the sense ‘know each other’, or a lexical (lexicalized) one, with the sense ‘be acquainted’, and the base verb znát ‘know’ can only be related to the former sense. The two senses cannot be easily distinguished in all contexts because of their proximity and the scarcity of contextual clues. Conversely, when there are contextual clues indicating single-event reciprocity, it may not be clear whether there is a lexicalized unit, which the contextual clues support, or whether the single-event construal arose in the context alone.

3. A sentence with a binary conjunctive reciprocal is synonymous with the conjunction of sentences containing the corresponding non-reciprocal predicates, e.g. Jan a Petr se navzájem navštívili ‘John and Peter visited each other’ = Jan navštívil Petra a (také) Petr navštívil Jana ‘John visited Peter and (also) Peter visited John’. Such conjunctions aptly illustrate the binary semantics of these reciprocals. Single-event (lexical) reciprocals do not allow such a transformation for the trivial reason that they have no base verbs, see the previous criterion. Even if there is a corresponding non-reflexive lexical unit, it may not be a true base verb, as suggested by the fact that a conjunction of sentences containing this verb is not an alternative to the sentence with the single-event reciprocal. For example, Jan a Petr se znají už dvacet let ‘John and Peter have known each other for twenty years’ cannot be replaced by Jan zná Petra už dvacet let a Petr zná Jana už dvacet let ‘John has known Peter for twenty years and Peter has known John for twenty years’ (this is not a communicationally acceptable alternative).

4. If there is a base verb, the reciprocal situation can also be expressed with the bipartite quantifier jeden druhý, e.g. navštívili se navzájem – navštívili jeden druhého ‘they visited each other’. Thus, this alternative also signals binary conjunctive reciprocity. However, this does not apply to reciprocals of the form hádali se jeden s druhým, where jeden s druhým alternates with spolu, see Section 4.

5. To render the reciprocal meaning, base verbs related to binary conjunctive reciprocals materialized by reflexive verbs can also be combined with the reflexive pronoun sebe (sobě). Actually, the replacement of the “light” marker se (si) by the “heavy” marker is obligatory in Czech: the long form must be selected when the reciprocal relation is conjoined or contrasted with a non-reciprocal one, e.g. Jan a Petr navštívili Martina i sebe navzájem ‘John and Peter visited Martin and also each other’; Jan a Petr navštívili Martina, sebe navzájem ale ne ‘John and Peter visited Martin but not each other’. As for single-event reciprocals, Czech speakers would probably not opt for the “heavy” marker, e.g. Jan a Petr se hádali spolu (^?mezi sebou) i s jinými lidmi ‘John and Peter argued with each other and also with other people’.

The situation in Czech is different from the situation in Polish, where, as Wiemer (1999) observes, the pronoun siebe replaces the clitic się not only for the syntactic reasons noted in the previous paragraph, but also for reasons related to the reflexivity-reciprocity homonymy of the reflexive forms. According to Wiemer (1999: 308), “a replacement of się by siebe most probably renders a reflexive reading with those lexical reciprocals that also most readily appear in any type of any comitative construction and with the reciprocity marker ze sobą […]. With less typical lexical reciprocals, which bear a more straightforward semantic relation to their transitive bases, siebie, on the contrary, causes a fragmentation of the denoted event and reinforces the reciprocal interpretation”. The first group comprises verbs such as widzieć się ‘meet’, całować się ‘kiss’, and (przy) witać się ‘exchange greetings’, whereas the second group includes verbs such as obejmować się ‘embrace each other’ and cmoknąć się ‘smack each other’. Wiemer notes that members of the latter group, unlike members of the former group, cannot occur with comitative objects. In Czech, the respective translation equivalents do not seem to form two groups on the grounds of the given criterion, since not only the constructions such as Jan se pozdravil s Petrem ‘John and Peter greeted each other’ but also the constructions such as Jan se objímal s Martinou ‘John and Martina were hugging each other’ are possible and unproblematic. Moreover, for any of the respective equivalents, a replacement of se (si) by sebe (sobě) cannot be licensed by the need to reinforce the reciprocal interpretation. If a reflexive reading is to be unambiguously expressed instead of the reciprocal (and more typical) one, a combination of the intensifier sám ‘by oneself’ with the pronoun sebe seems to be the preferred option, e.g. viděli v zrcadle sami sebe ‘they saw themselves in the mirror’. However, it is the intensifier rather than the “heavy” marker (or the combination of both) that evokes the reflexive reading, as suggested by the alternative viděli se sami v zrcadle ‘they saw themselves in the mirror’, which is also unambiguous. The fact that the opposition of se – sebe in Czech (and sa – seba in Slovak) does not follow the same principles as in Polish, has also been noted by Ivanová (2025).

1. In principle the possibility of a discontinuous form such as Martin se hádal s Petrem ‘Martin argued with Peter’ indicates lexical / single-event reciprocity. However, it can be debatable what exactly “to be possible” means. For example, the construction ^?Martin se respektoval s Petrem ‘Martin and Peter respected each other’ may be deemed ungrammatical by some speakers, but grammatical by others. It is not documented in the Czech National Corpus, but it can be found in Google.

2. Although this would actually be a petitio principii fallacy, I believe that the very opposition of the reciprocal adverbs navzájem/vzájemně and spolu is a good indicator of binary conjunctive and single-event reciprocity in Czech, respectively. This is a convenient criterion, because both adverbs, save for exceptions, adhere to reciprocals the reciprocants of which are expressed in the same syntactical position (typically: Jan a Petr se navzájem navštívili ‘John and Peter visited each other’, Jan a Petr se spolu pohádali ‘John and Peter had an argument’) and are thus fully symmetrical in meaning. The syntagmatic nature of this criterion also allows for its application in corpus-based research. However, even this criterion is not entirely unproblematic. First, navzájem/vzájemně “reliably” indicates binary conjunctive reciprocity only if the number of reciprocants equals two (see Section 6.2). Second, both markers sometimes co-occur (studenti spolu navzájem diskutovali ‘the students discussed with each other’). Third, some reciprocals do not follow the assumed pattern. For example, the reciprocal prolínat se ‘be intertwined’ is primarily combined with navzájem/vzájemně, despite being a single-event reciprocal, as suggested by application of criteria c)–f) (application of criteria a) and b) would be problematic). Similarly, the reciprocal lišit se (od sebe) ‘differ (from each other)’ can only be combined with navzájem/vzájemně, even though it is not a binary conjunctive reciprocal, as suggested by criteria b)–e) (application of criterion a) would be problematic; criterion f) fails here).

4 Morphological and lexical indicators of reciprocity in Czech

The means of expressing reciprocity in Czech are much more diverse than suggested in the previous section, since the morphological and lexical markers (reflexive morphemes and word forms, the expression jeden druhý and a few adverbial markers) occur and co-occur in multiple ways. In the following, I will identify several classes of reciprocals in Czech based on the number of markers used.^[12]

1. No marker:

Petr a Martin diskutovali ‘Peter and Martin discussed’, telefonovali ‘… talked on the phone’, bojovali ‘… fought’, etc. Without reciprocity markers and contextual clues, such expressions are vague, i.e., they can mean ‘… with each other’ as well as ‘… with someone else’. However, even with the latter sense, the situation described is reciprocal in nature.

1. A single marker:

1. The reflexive morphemes se, si and the “long” word forms sebe, sobě and sebou: nenáviděli se ‘they hated each other’, záviděli si manželky ‘they envied each other’s wife’, pohádali se ‘they argued’, nemohli bez sebe žít ‘they couldn’t live without each other’, mluvili o sobě hezky ‘they talked nicely about each other’, pohrdali sebou ‘they despised each other’, etc. Note that the reflexive morpheme is multifunctional, so a given predicate is often ambiguous, often between a reciprocal sense and a reflexive sense. For example, nenáviděli se can mean both ‘they hated each other’ and ‘they hated themselves’.

2. The expression jeden druhý ‘each other, one another’, whose second part can have different cases depending on the valency of the predicate:^[13] jeden druhého nenáviděl ‘one hated the other’/ nenáviděli jeden druhého ‘they hated each other’, jeden druhému záviděl manželku ‘one envied the other’s wife’ / záviděli manželku jeden druhému ‘they envied each other’s wife’, jeden druhým pohrdal ‘one despised the other’ / pohrdali jeden druhým ‘they despised each other’, etc. This marker is usually not used with the reflexive marker (*nenáviděli se jeden druhého), but the construction jeden s druhým ‘with one another’, which is rather a different marker, sometimes occurs with single-event reciprocals (věřící se můžou sejít jeden s druhým ‘the faithful can meet with each other’) and even with binary conjunctive reciprocals (nenáviděli se jeden s druhým ‘they hated each another’), where the reflexive morpheme is part of the sentence.^[14]

3. An adverbial marker. There are five of them in Czech: navzájem ‘each other, mutually’, vzájemně ‘each other, mutually’, spolu ‘with each other’, dohromady ‘together’, and mezi sebou ‘among themselves’.^[15] Being a single marker is quite typical for the marker spolu as well as for the (less frequent) markers dohromady and mezi sebou: bojovali spolu ‘they fought with each other’, bojovali mezi sebou ‘they fought with each other’, mluvili spolu ‘they talked to each other’, smíchat všechny ingredience dohromady ‘to mix all the ingredients together’. It is not very typical of the marker navzájem/vzájemně, although it is possible for some predicates, for example: studenti navzájem diskutovali ‘students discussed with each other’. In these cases, the adverbial marker can in principle be omitted, but this is not possible in all contexts, cf.: ten večer jsme (spolu) mluvili o všem možném ‘we talked about all sorts of things that night’ × už rok jsme spolu nemluvili ‘we haven’t talked for a year’. However, some single-event reciprocals require the spolu marker to be expressed if the verb is not used discontinuously (in which case the “s + N₇” complement is obligatory): soucítili spolu ‘they sympathized with each other’, spali spolu ‘they had sex’.

1. Two markers:

When a reflexive marker is used, an adverbial marker can be attached to the reciprocal. The reflexive marker can be considered a primary marker and the adverbial marker a secondary marker, since the presence of the latter is conditioned by the presence of the former, and the latter can in principle be omitted from the sentence: nenáviděli se (navzájem) ‘they hated each other’, potkali se (spolu) ‘they met’, etc. However, adverbial markers can also occur as primary markers, see type 2c) above.

This type also includes cases where two adverbial markers occur together without the predicate containing a reflexive component: navzájem spolu komunikují ‘they communicate with each other’.^[16]

1. Three markers:

Combinatorial possibilities in Czech are quite rich, although the combination of three markers cannot be considered typical. It is possible to find combinations such as navzájem se mezi sebou kontrolovali ‘they controlled each other’ and navzájem si spolu konkurují ‘they compete with each other’ (the latter example could also contain only two markers: navzájem spolu konkurují).

These formal types of reciprocals can be associated with the semantic types defined above as follows:

Binary conjunctive reciprocals always have at least one marker (nenáviděli jeden druhého, nenáviděli se ‘they hated each other’), while single-event reciprocals do not (diskutovali ‘they discussed’).^[17] Both single-event and binary conjunctive reciprocals can contain the reflexive morphemes se, si, while only the latter can contain the word forms sebe, sobě, sebou, excepting cases such as oddělit od sebe ‘separate from each other’, where the long form of the reflexive is required by a preposition. In other words, as part of reciprocal expressions, these forms are clear signals of binary conjunctive reciprocity; this does not apply to the adverbial marker mezi sebou, however.

In general, the use of the adverbial markers navzájem, vzájemně, spolu, dohromady, and mezi sebou is not directly conditioned by any semantic type, but it is related to syntactic patterns: It is only possible with reciprocals where the respective participants are expressed in a single syntactic position, typically the subject position, less often the object position, while it is not possible with discontinuous reciprocals where one of the participants is expressed with the form “s + N₇”, compare: Martin a Petr se spolu pohádali ‘Martin and Peter argued’– *Martin se spolu pohádal s Petrem ‘Martin argued with Peter’; Fotbalisté Sparty a Slavie se mezi sebou nesnášeli ‘The football players of Sparta and Slavia hated each other’ – *Fotbalisté Sparty se mezi sebou nesnášeli s fotbalisty Slavie ‘The football players of Sparta and Slavia hated each other’. As observed by Kettnerová and Lopatková (2020), the expression of spolu is possible only for reciprocals which can be used discontinuously, although spolu and the expression “s + N₇” cannot occur together, for example: Martin a Petr se (spolu) pohádali ‘Martin and Peter argued’ – Martin se (*spolu) pohádal s Petrem ‘Martin argued with Peter’.

5 Variation between reciprocals / reciprocal markers in Czech

The variation of the individual constructions is a phenomenon that is complex and difficult to explain. The variants differ from each other in the number of markers involved as well as in which marker is chosen from a given set of morphemic and lexical alternatives. I will give some examples of such competing constructions:

The marker jeden s druhým is an alternative to the adverb spolu: pohádali se spolu – pohádali se jeden s druhým ‘they argued one with another’. On the other hand, constructions containing navzájem/vzájemně are alternatives to constructions containing the marker jeden druhého, jeden druhému, etc., and lacking the reflexive morpheme: nesnášeli se navzájem – nesnášeli jeden druhého ‘they hated each other’. The marker mezi sebou seems to be compatible with both types, e.g. bavili se spolu / mezi sebou ‘they talked to each other’ (bavili se is a single-event reciprocal); navštěvovali se navzájem / mezi sebou ‘they visited each other’ (navštěvovali se is a binary conjunctive reciprocal).

Obviously, such a complex variation can hardly be treated in a single paper, unless it is to be described only superficially. For this reason, I have limited my attention to a pair of alternatives: reciprocals with the adverbial markers navzájem or vzájemně and reciprocals with the adverbial marker spolu. Since the adverbs navzájem and vzájemně are identical in meaning, distribution and stylistic value, I consider them to be variants of a single marker and represent them as navzájem/vzájemně. From the set of Czech adverbial markers of reciprocity, navzájem/vzájemně and spolu are of particular interest because of their connection to the semantic classes of single-event reciprocals and binary conjunctive reciprocals.

6 Navzájem/vzájemně and spolu used as non-basic markers of reciprocity

The insights presented in this section are based on the analysis of a) 20 most frequent collocations made up of a reciprocal containing the reflexive element se and the adverb navzájem/vzájemně, b) 20 most frequent collocations made up of a reciprocal containing the reflexive element se and the adverb spolu (see Veselý 2025), but they also concern some other (less frequent) reciprocals containing the morpheme se.

By default, navzájem/vzájemně indicates binary conjunctive reciprocity, and spolu indicates single-event reciprocity. When navzájem/vzájemně occurs with single-event reciprocals, and spolu occurs with binary conjunctive reciprocals, I consider this type of use to be non-basic. In these cases, a variation between the basic and the non-basic marker arises, because the non-basic marker can always be replaced by the basic one. The use of navzájem/vzájemně together with single-event reciprocals is facilitated when the number of reciprocants is greater than two (see below), whereas the use of spolu does not seem to be affected by this factor.

7 On the semantic relationship between navzájem/vzájemně and spolu

As suggested in the previous sections, the range of use of navzájem/vzájemně is certainly wider than that of spolu. This is manifested by at least two phenomena: 1. single-event reciprocity is less often marked with spolu than binary conjunctive reciprocity is marked with navzájem/vzájemně,^[35] 2. As regards the use of the non-basic variants, it is generally less common to combine a binary conjunctive reciprocal with spolu than to combine a single-event reciprocal with navzájem/vzájemně (the latter use is facilitated when the number of reciprocants is greater than two). It is certainly the adverb navzájem/vzájemně, not the adverb spolu (nor the adverbial expressions dohromady and mezi sebou, see Section 4), that can be considered the most typical adverbial marker of reciprocity in Czech. In view of the fact that in contexts with more than two reciprocants, the marker spolu can be replaced almost freely by the marker navzájem/vzájemně, the hypothesis could be put forward that the two markers form a privative opposition of lexical units with a grammatical function in which navzájem/vzájemně is the unmarked member, because it either signals binary conjunctive reciprocity or is indistinctive in terms of the opposition between binary conjunctive and single-event reciprocity (i.e., it only signals reciprocity in general), whereas spolu is the marked member, because, with few exceptions (see Section 6.1), it is distinctive in terms of this opposition, i.e., it signals single-event reciprocity. This hypothesis will be discussed in the remainder of this section. Despite the skepticism of some linguists (Haas 1957; Plungjan 1994; Haspelmath 2006 and others), the theory of markedness and the concept of the zero sign is still an influential theory; therefore, I think it is worthwhile to address this issue.

Originally, privative oppositions were identified in phonological systems (Trubeckoj 1939): members of the phonological correlation are unequal, i.e., one is marked, the other is unmarked (for example, in Czech /d/ and /t/ are voiced and voiceless phonemes, respectively, and the feature of voice represents the mark). However, Jakobson (1932, 1938) also applied this theory to morphological categories. Jakobson (1938) introduced the notion of the zero sign: while a marked sign signals A, a zero sign signals neither A nor not A. The zero sign is used in the proper sense (au propre), when A and not A are not distinguished, and when not A is to be signaled; however, the zero sign can also signal A.^[36] The opposition between “nothing” and “something” can be illustrated by an example given by Jakobson: osel ‘donkey’ – oslice ‘female donkey’ (the correlation of gender; the examples given are from Czech). While oslice necessarily refers to individuals of the female sex, i.e., it bears the feature A, osel is by default indifferent to the male-female opposition, although in some contexts, it can also refer only to individuals of the male sex, i.e., it bears the feature A’. Accordingly, osel is the unmarked member of this lexical pair, which is reflected by the absence of the suffix -ic(e).

Dokulil (1958) points out that a necessary condition for morphological correlation is the possibility of substitution of the marked member by the unmarked member provided that the respective feature is clearly indicated by the context. He gives three morphological oppositions in Czech that meet this condition: present conditional (přišel bych ‘I would come’) and past conditional (byl bych přišel ‘I would have come’), preterite (přišel jsem ‘I came’) and past preterite (byl jsem přišel ‘I had come’), iteratives (dělávat ‘do repeatedly’) and non-iteratives (dělat ‘do’). (It is not entirely clear to me whether the gender opposition would also pass this test in Dokulil’s view.) For example, the present conditional přišel bych can be used with the sense of the past conditional in a context where it is clear that the action may not happen (Kdyby venku tak strašně nepršelo, přišel bych / byl bych přišel ‘If it wasn’t raining so hard outside, I would have come’).

As regards the relationship between navzájem/vzájemně and spolu, the corpus data suggest that it is more common to use navzájem/vzájemně with single-event reciprocals (navzájem se hádat ‘argue with each other’) than to use spolu with binary conjunctive reciprocals (respektovat se spolu ‘respect each other’^[37]), even if contexts with more than two reciprocants are not considered. However, my analysis of the corpus data showed that the former use is far from common for most single-event reciprocals,^[38] which means that spolu as a marker of single-event reciprocity is not freely substitutable by navzájem/vzájemně, even when the single-event character of the action is clearly determined in a given context. I am therefore inclined to think that navzájem/vzájemně should not be treated as the unmarked member of the opposition between navzájem/vzájemně and spolu. I think it is more accurate to say that both adverbs are marked in their own way, i.e., each of them positively indicates a specific meaning: binary conjunctive reciprocity and single-event reciprocity, respectively. A similar type of opposition is assumed by Dokulil (1958) for the morphological category of number expressed by nouns: although Jakobson and some other members of the Prague school consider the singular to be the unmarked member of this opposition, Dokulil pointed out that in contexts like To je pěkná kniha ‘It is a nice book’, singular endings cannot be substituted by plural endings without a change in meaning. Such a substitution is possible only if the noun represents the whole class, as in the sentence Slavík pěkně zpívá ‘Nightingale sings beautifully’. This sentence has the variant Slavíci pěkně zpívají ‘The nightingales sing beautifully’, but in Dokulil’s opinion the two sentences are not completely identical in meaning: in the first sentence the noun denotes a species, whereas in the second sentence it denotes individuals (albeit all individuals of the respective species).

In principle, an opposition of linguistic units can be neutralized in two ways: a) it is not important for communication whether a certain characteristic applies to the denoted entity, or b) it is not (completely) possible to determine whether this characteristic applies due to a certain blurring of the opposition. The former type is represented by the category of gender. For example, the masculine noun tygr ‘tiger’ indicates the male sex in some contexts (V pražské zoo chovají tři tygry a dvě tygřice ‘The Prague Zoo breeds three tigers and two tigresses’), but it is usually used indifferently to the male/female opposition, because this feature is usually not important in terms of communication purposes (Děti viděli v zoo tygra ‘The kids saw a tiger in the zoo’). The latter type, on the other hand, can be represented by the above example Slavík pěkně zpívá ‘Nightingale sings beautifully’ / Slavíci pěkně zpívají ‘The nightingales sing beautifully’. In this case, the semantic opposition between singularity and plurality is somewhat blurred, since the subject noun denotes a single species consisting of many individuals. Arguably, the neutralization of the opposition between navzájem/vzájemně and spolu that occurs when a single-event reciprocal is used in a context with more than two reciprocants is also of this type: the reciprocal denotes a plurality of actions (provided it is an action predicate), and this plurality somewhat overshadows the single-event character of the reciprocal. Note that binary conjunctive reciprocals denote a plurality of actions in any case, regardless of the number of reciprocants, so the blurring assumed for the single-event reciprocals does not occur here.

However, the fuzziness between the two poles can also be a result of the nature of the reciprocal situation, and in this case it can also affect the binary conjunctive reciprocals. The reciprocals hecovat se ‘tease each other’, zdravit se ‘greet each other’, and objímat se ‘hug each other’ denote situations the parts of which (that can be expressed by the corresponding non-reflexive verbs hecovat, zdravit, and objímat, respectively) can also occur separately, i.e., independently of each other, but for which the bi-directional, reciprocal concept is typical. The close relationship between these parts is reinforced by the fact that they occur either simultaneously (objímat se ‘hug each other’) or in close succession (hecovat se ‘tease each other’, zdravit se ‘greet each other’). The borderline character of these reciprocals is indicated by the fact that they can be combined with both navzájem/vzájemně and spolu, see Section 6.1. Occurrences with navzájem/vzájemně predominated in the corpus data, but occurrences with spolu were relatively frequent, especially for the reciprocals objímat se ‘hug each other’ and hecovat se ‘tease each other’. Note that while the neutralization of the given opposition that occurs in situations with more than two reciprocants (see the previous paragraph) is a grammatical, systemic matter, the neutralization that results from the nature of the reciprocal situation, as in the case of the three reciprocals above, concerns only a number of specific units of the lexicon.

While the possibility of using the unmarked category in the sense of the marked category may be a necessary condition for understanding a given opposition as a privative one (see above), it is probably not true that only privative opposition are neutralized – equipollent oppositions, where both members positively indicate a certain grammatical feature, sometimes are too.

The equipollent character of the opposition between navzájem/vzájemně and spolu is also indicated by other facts: 1. Arguably, privative oppositions occur mostly in phonological and morphological systems; oppositions in lexical systems are usually equipollent oppositions (of course, privative morphological oppositions such as tiger × tigeress cannot be dissociated from the respective lexemes). 2. This is related to the fact that lexical signs, unlike morphological signs, cannot lack a phonetic form. Note that the opposition between “nothing” and “something” is traditionally assumed not only on the semantic level, but also on the formal level (see Křížková 1965): for example, in the case of the opposition between the present conditional and the past conditional in Czech, the former (unmarked) form lacks a component that the latter (marked) form contains; in other words, there is a correspondence between form and meaning. The adverbs navzájem/vzájemně and spolu are two “non-zero” forms, so it is not surprising that both have a certain positive meaning. Note also that the adverbs navzájem/vzájemně and spolu are not formally similar at all, and navzájem/vzájemně is not even shorter than spolu (a shorter form could be an indicator of an unmarked category). 3. The expression of both markers is usually only facultative, in contrast to the expression of morphological categories.

However, the non-basic use of both navzájem/vzájemně and spolu is not limited to the cases of neutralization. The difference between the two markers is that, arguably, spolu can only be combined with binary conjunctive reciprocals having certain semantic properties: Section 6.1 discussed cases in which a) the reciprocal doplňovat se ‘complement each other’ overlapped with the adverb spolu in terms of the semantic feature of cooperation, b) a certain blurring of the opposition between single-event and binary conjunctive reciprocity occurred (cases such as hecovat se ‘tease each other’ and zdravit se ‘greet each other’), c) the meaning of the reciprocal navštěvovat se was shifted to the pole of single-event reciprocity. Navzájem/vzájemně, on the other hand, does not require the predicate to have such properties. Let me return to two examples mentioned in previous sections:

The reciprocal navštěvovat se ‘visit each other’ probably cannot be said to denote a situation the parts of which are not easily discernible, because the individual events (i.e., visits) do not overlap in time nor do they need to occur in close succession. However, the reciprocal navštěvovat se in (44) can be assumed to be shifted to the pole of single-event reciprocity, since it is close in meaning to the verb setkávat se ‘meet (on a regular basis)’ (which is a single-event reciprocal). The occurrence of navštěvovat se with the marker spolu is a consequence (and an indicator) of this shift. The reciprocal rozejít se ‘break up’, on the other hand, when all the reciprocants are expressed in the subject position, denotes a single, fully symmetrical event. This also applies to example (45): arguably, the binary “tune” of the reciprocal situation is not encoded in the lexeme rozejít se, but it is alluded to by the marker vzájemně. In fact, the adverb is used “in defiance” of the lexical meaning of the reciprocal, forcing a certain modification of its meaning. Navzájem/vzájemně may not be an unmarked member of a privative opposition (see the discussion above), but of the two reciprocity markers it is certainly the member with the wider range of application: even if it does not coincide with the meaning of the verbal lexeme, it can sometimes be combined with it. In these cases, a certain “semantic tension” may be perceived, resulting from the discordance between the meaning of the reciprocal and the meaning of the marker.^[39] Significantly, such occurrences are rather occasional, in contrast to the use of navzájem/vzájemně with single-event reciprocals made possible by the number of reciprocants being greater than two (where, arguably, no “semantic tension” arises). Since the opposition between the markers navzájem/vzájemně and spolu is equipollent (see above), occurrences such as (45) can be reminiscent of cases when a marked member of a morphological opposition is replaced by another marked member of the same opposition, e.g., when the second singular is replaced by the first plural (Kampak jdeme, panáčku? ‘Where are we going, dummy?’). In such cases, the semantic effect of the transposition is always strong.

8 Conclusions

In most cases, both the markers navzájem/vzájemně and spolu are used as secondary markers of reciprocity, i.e., reciprocity has a double marking in the sentence. However, the distinction between single-event and binary conjunctive reciprocity is not indicated with the reflexive morphemes se and si, which may be the reason for the use of the lexical markers. This is not to say that navzájem/vzájemně and spolu always really indicate one of the two types. When they become variants, the non-basic variant cannot be said to mark the semantic type which it marks by default. Whether it is a systemic variation conditioned by the number of reciprocants being greater than two (žáci ve třídě se spolu/vzájemně hádali ‘pupils argued with each other in the class’) or an occasional variation made possible by the lexical meaning of an individual reciprocal (hráči se vzájemně/spolu hecovali ‘the players teased each other’; vzájemně se navštěvovali / navštěvovali se spolu ‘they visited each other’), the non-basic variant (always given in second place) does not signal that hádat se ‘argue’ is a binary conjunctive reciprocal, and that hecovat se ‘tease each other’ and navštěvovat se ‘visit each other’ are single-event reciprocals, respectively (since they are not, although hecovat se and navštěvovat se are somewhat shifted to the pole of single-event reciprocity). Of course, they also do not signal the opposite semantic types (associated with the basic variants), because this would contradict their lexical meaning. Thus, both markers used as non-basic variants only indicate reciprocity in general, but they are by no means agnostic in terms of the given semantic opposition: there is usually a “good reason” for the use of the non-basic variant, i.e., the situation denoted must have features that motivate / enable the speaker’s choice of this variant, and which in turn are suggested to the perceiver by the non-basic variant. So, it is the nature of the reciprocal situation expressed by the reciprocal which motivates the speaker to use the non-basic variant, and which also makes this variant semantically defective in the above sense. It can be debated whether (and to what extent) the non-basic variant also semantically affects the reciprocal. I think that it is always indicative of the nature of the situation, and only sometimes, in cases like rozešli jsme se vzájemně ‘we broke up with each other mutually’ (see example (45) above), does it add semantic features to the sentence that would not otherwise have been associated with it.

Several types of variation between the markers navzájem/vzájemně and spolu can be distinguished, and these types are reflected by the relative frequencies of the respective variants. In the case of the systemic variation which is conditioned by the number of reciprocants being greater than two, the variation occurs with a number of reciprocals (although in varying degrees), and the frequencies of the non-basic variants are relatively significant (when compared to the basic variants). On the other hand, the variation made possible by indistinctness of the denoted situation in terms of the opposition between single-event and binary conjunctive reciprocity concerns only a few reciprocals, and it is the case that frequencies of the non-basic variants are less significant (e.g. hecovat se spolu ‘to tease each other’). Finally, when navzájem/vzájemně used as a non-basic variant modifies the meaning of the sentence, pointing out that two reciprocants are active independently of each other, the frequencies of the non-basic variants are low (e.g. radit se navzájem ‘consult each other’), unless, for example, the given syntactic construction becomes a cliché (dohodnout se vzájemně na ukončení smlouvy ‘to mutually agree to terminate the contract’).^[40]

The use of an adverbial marker of reciprocity has different motives as well as semantic effects depending on whether it corresponds to the semantic type of the reciprocal (in terms of the distinction between the single-event and binary conjunctive reciprocity), but also depending on the characteristics of the individual reciprocals, i.e., their lexical semantics, systemic relations in the lexicon, etc. Nevertheless, it is possible to assume that the respective markers have consistent semantic effects in all contexts, albeit to varying degrees. These effects can be described as follows: Navzájem/vzájemně indicates that the reciprocal situation consists of events (in the broadest sense) that are independent on each other and easily distinguishable, and that the participants engaged in this situation switch across the individual events. Spolu, on the other hand, indicates that the reciprocal situation consists of events (or rather sub-events) that are not independent of each other and easily distinguishable, and that the participants jointly bring about the situation, i.e., they participate in the same way. Note that the function of navzájem/vzájemně here is conceived in such a way that it covers both the cases where it is combined with binary conjunctive reciprocals and the switching of participants equals to switching of semantic roles, and the cases where it is combined with single-event reciprocals and the switching of participants means that the individual events (as parts of the reciprocal situation) involve different participants who, however, fulfill only a single semantic role. Compare e.g.: A a B se navzájem nenáviděli ‘A and B hated each other’ = A hated B and B hated A × týmy A, B a C se navzájem utkali ‘teams A, B and C fought each other’ = A and B fought each other, A and C fought each other, B and C fought each other. Both adverbs have in common that when they are used as non-basic markers of reciprocity, they often point out the active role of the individual participants (but each in a different way), e.g. trenér a klub se vzájemně dohodli na ukončení spolupráce ‘the coach and the club mutually agreed to end their cooperation’; oba hráči se spolu skvěle doplňovali ‘both players complemented each other perfectly’.