nueva página del texto (beta)
Inglés (pdf)
Artículo en XML
Referencias del artículo
Traducción automática
Enviar artículo por email
Citado por SciELO 
Similares en
SciELO 
Permalink
1. Introduction
Yucatec Maya (the Mayan language spoken in the yucatan Peninsula, Mexico, and parts of Belize; ISO code: yua), like other Mayan languages, shows movement of a focused constituent to the left edge of the clause, as illustrated in the following examples:1
It has been widely debated in the literature on focus in Mayan languages whether these constructions are monoclausal focus constructions (like those of Hungarian, for instance: Kiss 1998) or clefts. In a recent paper, however, Verhoeven & Skopeteas (2015) present strong evidence that focus constructions in Yucatec like (1) and (2) are not clefts. The gist of the argumentation presented by these authors is that (a) there is a different set of focus constructions that are unquestionably clefts, and (b) these cleft constructions behave differently from focus constructions like (1) and (2). Following this line of argumentation, in this paper I present a new set of data that further supports the conclusion that focus constructions like the ones above are not clefts. Along with the diagnostics originally identified in Verhoeven & Skopeteas (2015), the ones I propose here add up to a considerable battery of tests that can be applied to focus constructions in other Mayan languages in order to determine whether they constitute ordinary monoclausal focus constructions or clefts.
Before proceeding any further, it is necessary to address an important terminological issue.2 Simplifying somewhat, in purely descriptive and typological terms a cleft is a biclausal copular construction consisting minimally of an informationally relevant constituent (the clefted constituent) and some form of reduced relative clause (the extra-focal clause). Crosslinguistic variation with respect to this minimal structure, of course, is widely reported in the literature. For instance, besides these two elements, English clefts show an overt copula and an impersonal pronoun it; clefts in Spanish have an overt copula but no impersonal pronoun, etc. In the theoretical literature, though, the term cleft has a much more specific meaning, namely, clefts are understood as structures of this type, but which furthermore show a number of characteristic properties not attested in other copular constructions (the best known of which are connectivity effects with respect to anaphor binding; see Reeve 2012: chapter 2 for a recent overview of these properties and relevant references). There is yet no description or analysis of these fine-grained properties in the constructions that are labeled as clefts throughout the literature on Yucatec, and it is not the purpose of this paper to provide such a description and analysis. As such, in order to highlight that there is no implication that the Yucatec structures analyzed here have any or all of the characteristic properties of clefts in English and other familiar languages, in what follows I use the cover term cleft construction, and not cleft, to refer to these constructions.
The same point can be made about the term pseudo-cleft. In purely descriptive terms, a pseudo-cleft can be defined as a copular construction where one of the arguments of the copula is (typically) a wh-clause (as in den Dikken 2006: 309-310). However, in the theoretical literature the term pseudo-cleft is most often used to refer specifically to specificational pseudo-clefts (broadly speaking, pseudo-clefts where the clefted constituent characteristically is a focus).3 The distinction is important because, as is well known, specificational pseudo-clefts have a number of properties that make them different from similar copular constructions (i.e. again, connectivity effects with respect to binding of anaphora, case, etc.; see den Dikken 2006 for a thorough survey of these properties and relevant references). In what follows, I provide a description and analysis of a number of cleft constructions in Yucatec that on the first approximation would appear to be pseudo-clefts. However, just as with regular cleft constructions, at this point I make no claim that these constructions have the properties characteristic of pseudo-clefts in English, Spanish, and related languages. In order to refer to these constructions without any implication about the presence of these properties, I use the cover term wh-cleft construction instead of pseudo-cleft. Future research will be necessary to determine whether the term pseudo-cleft is an appropriate label or not for the Yucatec wh-cleft constructions analyzed in this paper.
Having made this terminological clarification, the remainder of the paper is organized as follows: in section §2, I describe the basic properties of focus constructions in Yucatec, and I present a brief outline of their two possible analyses (i.e. as monoclausal constructions or as biclausal cleft constructions). In section §3, I provide a basic description of cleft constructions in Yucatec, including wh-cleft constructions, which have not been previously described for this language but which play an important role in my argumentation. In section §4, I present the comparison between cleft constructions and ordinary focus constructions, which points to the conclusion that ordinary focus constructions are not cleft constructions, and in section §5, I present my conclusions.
2. Focus constructions in Yucatec
2.1 Preliminary description
Yucatec Maya is a strictly head-marking language that displays mostly nominative-accusative syntax, but has an ergative-absolutive cross-referencing system that is split on the basis of mood and aspect, as is well known. Most clauses in Yucatec consist minimally of the verb and a proclitic (glossed ERG) cross-referencing the transitive subject of the verb. The main verb in turn displays a series of suffixes (glossed ABS) that agree with the object, and also with intransitive subjects in the moods and aspects where the ergative-absolutive cross-referencing system is observed. It should be noted that, because of split ergativity, not every instantiation of a pronominal element labeled as erg and abs necessarily bears an ergative or absolutive grammatical relation. Most of the time, the ergative proclitic is also preceded by an auxiliary. This minimal structure is shown in (3).
Argumental phrases appear to the left and right of this basic structure. The precise characterization of word order in Yucatec, however, is still an ongoing debate. It is widely agreed that the unmarked word order of intransitive clauses is VS, but whereas some works describe the basic word order of transitive clauses as VOS (Skopeteas & Verhoeven 2005; 2009b; 2009a; 2012), other works argue that SVO is the unmarked word order in this case (Durbin & Ojeda 1978; Hofling 1984; Briceño Chel 2002; Gutiérrez-Bravo & Monforte 2010). Here I adopt the latter analysis; this assumption does not have any consequences for the analysis of focus that follows.
2.2 Focus constructions
As mentioned, focus constructions in Yucatec have been widely analyzed in the literature, as in Bricker (1979), Bohnemeyer (2002), Tonhauser (2003), Verhoeven (2007), Skopeteas & Verhoeven (2012), Lehmann (2008), Gutiérrez-Bravo & Monforte (2011), AnderBois (2012), Vapnarsky (2013), Verhoeven & Skopeteas (2015), to mention just the most important references. Here I only present a basic description of this phenomenon. The two properties of focus constructions that will be crucial for the analysis presented in what follows are that (a) narrow contrastive foci surface in the position immediately to the left of the auxiliary (or immediately to the left of the verb, when there is no auxiliary),4 and (b) when the focus is a transitive subject, a special form of the verb (known in Mayan linguistics as agent focus) is observed. The following examples illustrate these two properties. In (4), the direct object is displaced from its canonical post-verbal position to the position immediately to the left of the auxiliary. In (5), focus on the transitive subject triggers the agent focus form of the verb, which in Yucatec is mostly characterized by the absence of both an auxiliary and the ergative proclitic cross-referencing the subject. For further details on the agent focus construction in Yucatec, I refer the reader to the texts cited above.
As mentioned in the introduction, it has been intensely debated whether these constructions are monoclausal focus constructions or cleft constructions. The proponents of the monoclausal analysis (Lehmann 1998; Skopeteas & Verhoeven 2012; Gutiérrez-Bravo & Monforte 2011) propose that there is a specific (and unique) position inside the clause which is reserved for the focus when there is one. These analyses very often follow the monoclausal analysis of Tzotzil focus constructions in Aissen (1992), whose defining characteristic is that the focus occupies the specifier position of the functional projection that is immediately above the verbal layer of the clause (see, for instance, Skopeteas & Verhoeven 2012). In what follows, I adopt the basics of this analysis, although with a more elaborate functional layer than the one in Aissen (1992).
My basic assumptions about the clause structure of these examples are the following, where I use example (2), repeated here as (6a), for illustration: (a) the aspect auxiliary (yaan, k, etc.) is the head of an aspect/mood phrase AspM-P, which is basically equivalent to TP in other languages (Yucatec is a tenseless language, as is well known: see for instance Bohnemeyer 2002); (b) the ergative proclitic that is coreferential with the subject is the head of a Subject Agreement Phrase (AgrS-P); (c) there is no verb movement in Yucatec, so the verb always surfaces inside the VP: vP is omitted for brevity:
The alternative to this analysis is to consider focus constructions to be cleft constructions (Bricker 1979; Bohnemeyer 2002: 116-129; Bohnemeyer 2009; Tonhauser 2003; Vapnarsky 2013). In these analyses, the focused constituent is taken to be a predicate, and the material that appears to the right of the focus is taken to be the extra-focal clause characteristic of cleft constructions. For the most part, these analyses are not explicit about further details of the syntax of these constructions, with the exception of Tonhauser (2003), who proposes the structure in (7a). For the purposes of the discussion that follows, here I assume the nullcopula structure in (7b) to be the predicative structure characteristic of cleft constructions. I myself adopt a similar (albeit more detailed) structure for true cleft constructions in (29).5
2.3. Focus constructions versus cleft constructions
Recently, new light has been shed on this debate by the analysis in Verhoeven & Skopeteas (2015). This analysis aims at solving the debate regarding the nature of focus constructions in Yucatec by looking at constructions that are uncontroversially cleft constructions and comparing them with the focus constructions illustrated above. The result is that focus constructions are different from true cleft constructions with respect to the following four properties: (a) the focus must show agreement with the verb in ordinary focus constructions (modulo agent focus), but not in cleft constructions; (b) a focused subject binding a reflexive pronoun must have the exact same features as the reflexive pronoun in ordinary focus constructions but not in cleft constructions; (c) the interrogative particle wáaj which was previously assumed to be able to attach only to predicates (i.e. Tonhauser 2003) such as a clefted argument (and foci like those in 4 and 5), is shown not to be a predicate-final particle; and (d) the focused XP of regular focus constructions is compatible with additive and scalar focus operators (xan 'also', and tak 'even', respectively: cf. Krifka 2007), but the focused XP of cleft constructions is not. Summing up, very specific syntactic evidence is provided by these authors that points to the conclusion that regular focus constructions are not cleft constructions in this Mayan language.
Taking the central idea in Verhoeven & Skopeteas (2015) as a starting point, in this paper I appeal to the same methodology, i.e. constructions that are uncontroversially biclausal cleft constructions are compared with ordinary focus constructions to see if they display other differences apart from those already identified by these authors. The results are presented in what follows, where I provide evidence that there are four further differences between ordinary focus constructions and cleft constructions: consequently, my analysis provides further support for the monoclausal analysis of the focus constructions illustrated so far.
3. Cleft constructions in Yucatec Maya
3.1. Preliminary description
Although some observations about the properties of cleft constructions in Yucatec have been made in works like Tonhauser (2003), Skopeteas & Verhoeven (2012), and Verhoeven & Skopeteas (2015), as previously mentioned there is still no detailed description of cleft constructions in this language. In this section I attempt to provide a starting point for such a description by giving a general overview of the properties of cleft constructions in Yucatec. Yucatec does not have a copula, and so (pre-theoretically, at least) cleft constructions in this language can be understood as a clefted constituent that appears with a headless relative clause (rc: the extra-focal clause). To illustrate the properties of these headless relative constructions, it is worth looking briefly at the properties of headed relative clauses. Yucatec has two basic kinds of relative clause, which in typological terminology correspond to gap relative clauses, and pronominal relative clauses, which show a fronted relative pronoun (máax 'who', ba'ax 'what', tu'ux 'where', or bix 'how'). These two kinds of relative clause are exemplified in (8) and (9), respectively. In (8), the underscore represents the canonical position of the relativized argument.
As can be seen in these examples, relative clauses in Yucatec are never introduced by a complementizer, and they do not have any morphology that makes them different from matrix or other subordinate clauses. However, in many cases they can be identified as relative clauses because they are bracketed by the deictic clitics (=o', in these examples) which obligatorily appear in the presence of the demonstrative le, and which attach to the right edge of the nominal expression.6 As such, the presence of these clitics in the examples above indicates that the relative clause is embedded inside the DP headed by le.
Simplifying somewhat, there are basically two different kinds of headless relatives in Yucatec. In the first kind, the head noun characteristic of headed relatives is absent, but the headless relative still shows the demonstrative determiner le (as in 10 and 11). In the second kind of headless relative, both the head noun and all the possible pronominal modifiers are absent altogether. In this latter case, when the corresponding relative clause is a pronominal relative clause, the result is a free relative clause, as in (12).7 In these examples the underscore corresponds to the position of the omitted nominal head.
Returning now to the specific properties of cleft constructions in Yucatec, these are built with an XP that typically functions as a focus plus any of the different kinds of headless relatives (i.e. the extra-focal clause) illustrated above.8 This is shown in the following examples. In (13) the headless relative clause of the cleft is a gap relative clause, whereas in (14) the headless relative is a pronominal relative clause. Observe that in (13) the verb of the headless relative shows agreement with the clefted constituent (the second person pronoun), and not default third person agreement.9 This agreement pattern (which may possibly be considered a particular type of connectivity effect) would appear to be some-what unusual, but has also been reported for numerous kinds of clefts in European and Brazilian Portuguese (Kato & Ribeiro 2009: 131-133).
From the perspective of a language like English it might seem odd that the extra-focal clause is introduced by a determiner, but this is (for instance) what is observed in Spanish clefts too, with the difference that Spanish has a copula and Yucatec does not.
Yucatec cleft constructions where the headless relative is a gap relative like (13) are described in Tonhauser (2003) and Verhoeven & Skopeteas (2015). Cleft constructions where the headless relative is a pronominal relative clause (i.e. wh-cleft constructions) have not yet been described for this language, an issue which I address in the following section.
3.2. Wh-cleft constructions
Following Baker (1989), Heycock & Kroch (1999), Huddleston & Pullum (2005), and Hartmann & Veenstra (2013), I take wh-cleft constructions to be cleft constructions where the headless relative that functions as the extra-focal clause is a headless relative with a wh-pronoun, and not a gap relative, as in the following examples of pseudo- clefts from English (16) and Spanish (17).
Definitions of English pseudo-clefts usually also further specify that pseudo-clefts lack the expletive pronominal it characteristic of it-clefts, so the English translation of the example (17) is not an English pseudo-cleft in these definitions.10 However, this specific part of the definition of wh-cleft constructions is not relevant for Yucatec, which lacks expletives altogether. Wh-cleft constructions are readily observed in Yucatec, and (14) in fact is an example of this kind of cleft construction. More examples of Yucatec wh-cleft constructions are presented below.
Just like regular cleft constructions (Verhoeven & Skopeteas 2015), wh-cleft constructions have the unmarked word order focus-[extra-focal clause], which can be observed in the examples above. Also, as can be seen with these examples, the headless relative functioning as the extrafocal clause can be either a free relative, or a pronominal headless relative introduced by le. However, elicitation shows that wh-cleft constructions with free relatives like (18) have a more restricted distribution. In the following examples it is shown that speakers mostly reject the equivalents of (18) and (20) if the headless relative is not introduced by le.
As noted by a reviewer, the unacceptability of the examples above might be directly related to the informational structure properties of the wh-cleft construction as a whole. In the standard case, the wh- clause of wh-cleft construction expresses old information (den Dikken 2006), or (more precisely) presupposed material (Prince 1978; see also Heycock & Kroch 2002: 149-150). When the wh-clause is introduced by the demonstrative le, as in (19) and (20), it is transparent that it corresponds to presupposed information. It may then be the case that the absence of le makes the corresponding constructions infelicitous unless the presupposed nature of the wh-clause is provided by the preceding discourse, which is presumably what happens in the text example in (18).
Now, a property regularly found in (specificational) wh-cleft constructions is their word order flexibility. For instance, English pseudo- clefts can show an inverted or reversed order, where the focus precedes the extra-focal clause, as in (23).11
Wh-cleft constructions in Yucatec also have this property. In Yucatec reversed wh-cleft constructions, the headless relative comes first and the focus follows. In texts, the focus of reversed wh-cleft constructions is typically a complete clause, as in (24), but examples can be found in which the focus is a smaller XP, like the NP in (25).
Observe that the clefted constituent in these cases is still a focus, which is different from what is characteristically observed in English reversed pseudo-clefts (Heycock & Kroch 2002). Now, just like regular wh-cleft constructions, reversed wh-cleft constructions have a more limited distribution and in elicitation they are often not accepted by speakers when the headless relative clause is not introduced by the demonstrative le, although speakers do accept some examples like (28). Again, it is likely that the elicited examples without le are not acceptable because they fail to signal the wh-clause as presupposed information.12
Before I present the evidence that indicates that regular focus constructions are not clefts I present my assumptions regarding the structure of both regular cleft constructions and wh-cleft constructions. I assume that both kinds of cleft construction are copular constructions. Following the analysis of Yucatec copular constructions in Armstrong (2010), I adopt an analysis where these constructions correspond to a Predicative Phrase (PredP). Just as in Armstrong's analysis, I assume that PredP projects its specifier to the right. I further assume that the focus XP of the cleft construction is externally merged (i.e. base-generated) in this specifier position, and that the headless relative is the complement of the head Pred. More importantly, I also assume that there is a functional projection above the PredP, which I provisionally assume to be the A/M-P of (6). The resulting structure (before any movement has taken place) is illustrated below.
For both kinds of clefts, the extra-functional layer in the structure proposed above allows for a straightforward analysis where, in the unmarked word order focus-[extra-focal clause], the constituent that is in the specifier position of PredP moves and internally merges as the specifier of A/M-P, as in (13-14) and (18-20). In the case of reversed cleft constructions and reversed wh- cleft constructions, I simply assume that there is no movement of the focused XP. Since this XP is base-generated in the specifier position of PredP, this automatically results in the reversed order [extra-focal clause]-focus observed in (24-25) and (28), where the focus is to the right of the extra focal clause.
There is actually an alternative way to derive the [extra-focal clause]- focus word order.13 This word order can also be derived in an analysis where first the focus moves from [Spec, Pred] to [Spec, A/M], and then the extra-focal clause moves to a dislocated/left peripheral topic position higher than A/M-P.14 This analysis provides a unified position for focus constituents ([Spec, A/M]), and might also account (in the absence of another deictic clitic) for the presence of the topic marker -e' in the right edge of the extra-focal clause in (24) and (28). However, this alternative analysis is a case of string-vacuous movement. As such, in the absence of further evidence of this string-vacuous movement, I do not adopt this analysis in what follows, and instead I adopt the simpler analysis laid out above, where the extra-focal clause always remains in its base-generated position, and where the word order differences observed depend exclusively on whether the focused XP moves to [Spec, A/M] or not.
4. Focus constructions are not cleft constructions
4.1. Agent focus
As mentioned in §2.2, when the focus is a transitive subject, a special form of the verb is observed. In (30) and (31) focus on the transitive subject triggers agent focus, which, as mentioned, in Yucatec is basically characterized by the absence of both an auxiliary and the ergative proclitic cross-referencing the subject.
This same property is observed in cleft constructions, as illustrated by the following examples, where it can be seen that there is no auxiliary or ergative proclitic to the right of the demonstrative le. Observe that it is possible to tell that (33) is an agent focus construction (where the -ej suffix has been dropped) because it is a morphologically transitive verb form, as evidenced by the presence of the present perfect suffix -m, which only appears in transitive constructions (see Bricker et al. 1998).
But in contrast with regular focus constructions, the agent focus form of the verb is not obligatory in cleft constructions when the focus is a transitive subject. Examples where cleft constructions do not show agent focus are readily found in texts, and in fact the absence of agent focus is also observed in the examples of cleft constructions presented in the elicited data in Verhoeven & Skopeteas (2015), of which (36) is an illustrative example.15
This difference is straightforwardly accounted for in the analysis where regular focus constructions are monoclausal, in contrast with true cleft constructions, which are biclausal. This is because the extra-focal clause of cleft constructions is a headless relative, and relativization of a transitive subject can, but need not trigger agent focus in Yucatec, as has been observed in Bricker (1979) and Gutiérrez-Bravo & Monforte (2011), and experimentally confirmed in Norcliffe & Jaeger (2016). This is illustrated in the following example where the transitive subject of the relative clause (i.e. 'a huaya tree') has been relativized, but the main verb of the relative clause does not show agent focus, as is apparent because of the presence of both the habitual auxiliary and the ergative proclitic. Example (8) above also shows this property.
4.2. Inversion
In the description of wh-cleft constructions in the previous section it was already mentioned that they can show inversion, which results in a reversed wh-cleft construction. Regular cleft constructions can also undergo inversion, as observed in Verhoeven & Skopeteas (2015), and so inversion can be taken to be a general property of cleft constructions in Yucatec. Examples of both kinds of reversed cleft constructions are provided below.16
In contrast, inversion in regular focus constructions is robustly ungrammatical.
It could be argued that the ungrammaticality of (41b) might be independently accounted for if verbs showing agent focus morphology require (for some reason) the focus to appear immediately to their left. However, this possibility is hard to reconcile with the data in (32) and (33), where the extra-focal clause equally shows agent focus morphology, but the element to its left is not the focus but the demonstrative determiner le.17
4.3. Negative pronouns
The third piece of evidence that regular focus constructions are not cleft constructions comes from negative pronouns. Yucatec is a negative concord language where negative pronouns can appear in the post-verbal field in which case they need to be licensed by one of various forms of sentential negation (see Gutiérrez-Bravo 2015a).
Alternatively, these negative pronouns can appear in the preverbal position, as in (43) and (44), in which case no sentential negation is observed. There is some evidence that in these cases the fronted negative pronoun occupies the focus position since, as noted originally in Bricker (1979), when the negative pronoun corresponds to a transitive subject it triggers agent focus morphology on the verb, as in (44).
As it happens, true cleft constructions are markedly different since they do not allow the focus to be a negative pronoun: this is in fact a common property of cleft constructions, since cleft constructions include an inherent presupposition of existence (Percus 1997; Geurts & Sandt 2002; Abusch 2010) whereas ordinary focus constructions do not (see Hartmann & Veenstra 2013 for a survey). The following elicited minimal pairs illustrate this.
It should be pointed out that in some cases speakers do accept constructions like the ungrammatical examples above. Example (49) is one such case, where "Mr. Nobody" means that the person under consideration is an insignificant individual.
However, it is clear that in this case mixmáak 'nobody' is not functioning as a negative pronoun: this is because, in contrast with true negative constructions, part of the meaning of (49) is clearly ∃x & exit' (x), i.e. the presupposition of existence mentioned above. Native speakers actually have very clear intuitions about this and have no problem pointing out that these constructions are not equivalent in meaning to constructions with true negative pronouns like (45) and (47).
4.4. Verb focus
The fourth and final piece of evidence that focus constructions are not cleft constructions comes from Yucatec verb focus constructions. In verb focus constructions the focused verb also surfaces in the left edge of the inflectional layer of the clause, but in this case a light verb beet/meen-t 'do' is inserted as the main verb of the construction (Bohnemeyer 2002; Lehmann 2008; Sobrino 2010).
As discussed in detail in Gutiérrez-Bravo (2015b), in these constructions the arguments of the focused verb (including the patient/direct object) show agreement with the light verb and not with the focused verb. This pattern is most clearly observed when the focused verb is a transitive verb. In this case, the focused verb appears in the morphological form resembling a gerund, and the patient of the transitive verb is expressed morphologically on the light verb, as in (51). This is not immediately obvious when the patient is 3rd person singular, as in (51) and (52), since the 3rd person singular absolutive suffix is -ø. However, this property of verb focus constructions becomes apparent when the patient corresponds to any other grammatical person, as shown in (53) and (54).
The patient expressed morphologically on the light verb is a full syntactic direct object, as can be observed in the fact that it surfaces as a subject if the light verb is passivized, in which case it is expressed by an ergative pronoun in constructions with a nom-acc alignment (indicative mood) or with an absolutive suffix in constructions with an erg-abs alignment (perfective and irrealis mood).
Now, the relevant point for our discussion is that it is not possible in Yucatec to construct cleft constructions equivalent to the verb focus constructions illustrated so far. This is shown in the following elicited examples where the extra-focal clause after the focused verb is no different from the headless relatives characteristic of other kinds of cleft constructions.
It might be argued that constructions like (58) and (60) could be independently ungrammatical because of some semantic restriction against having the extra-focal clause introduced by a determiner in cleft constructions when the focus is a verb. However, this alternative analysis would find it hard to explain why such a semantic restriction is not present in languages like Spanish, which do allow the extra-focal clause of verb cleft constructions to be introduced by a determiner, as shown in the examples below.
Furthermore, in the analysis I have adopted here there is a straight- forward way to understand the ungrammaticality of (58) and (60). In regular verb focus constructions, the semantic patient of the focused verb is not morphologically realized on the focused verb itself, but it is nonetheless realized in the same clause. In contrast, in the biclausal analysis of cleft constructions I have defended here, the patient is realized inside the extra-focal clause, but the focused verb in (58) and (60) is not a part of this clause to begin with. Although it is beyond the scope of this paper to provide a detailed analysis of the specific locality conditions that disallow cleft constructions where the verb is the focus, the relevant point is that once more we observe a specific difference between regular focus constructions and cleft constructions, which supports the conclusion that both constructions are distinct in Yucatec, and hence that regular focus constructions are not cleft constructions.
5. Conclusions
In this paper I have addressed the ongoing debate in Mayan linguistics regarding whether or not focus constructions in Yucatec Maya are regular, monoclausal focus constructions or cleft constructions. I have provided four new types of evidence that focus constructions are not cleft constructions. Specifically, I have presented evidence that focus constructions differ from true cleft constructions with respect to four properties: (i) the obligatory nature of agent focus morphology; (ii) the possibility of inversion; (iii) the possibility of having a negative pronoun as the focus, and (iv) the possibility of having verb focus constructions. Along with the diagnostics originally identified in Verhoeven & Skopeteas (2015), the ones I have proposed here add up to constitute a considerable battery of tests that can in principle be applied to focus constructions in other Mayan languages, in order to determine whether they constitute ordinary monoclausal focus constructions or cleft constructions.
