DOES NOT EXIST

RRL, LXIV, 2, p. 177–188, Bucureşti, 2019

DOES NOT EXIST

ANNA SNARSKA

Abstract. This paper is a contribution to the debate about the untractable nature of the phenomenon of partial control (PC). Analyzing data from Polish, I show, pace Landau (2000), that PC is not limited in its occurrence by the semantic type of the matrix verb. Furthermore, I argue that the availability of the ‘group’ interpretation of the non-lexical subject is not contingent upon the semantic type of the verb but is determined by the semantics of the verb in a lower clause. Only inherently reciprocal verbs like meet or argue are capable of generating the meaning in question. These verbs, taking a null discontinuous phrase, introduce an additional participant of the act denoted by the verb; hence the alleged ‘PRO1+ effect’. This entails that PC, welcomed in the linguistic world as a new subtype of control, is simply a phantom.

Keywords: control, PRO, reciprocal verbs, collective verbs.

1. INTRODUCTION

Many accounts have been offered to explicate an enigmatic nature of partial control (henceforth PC), a phenomenon that initially had gone unnoticed until Landau (2000) brought it to the attention of the linguistic world. This proliferation of analyses notwithstanding (cf. Martin 1996; Hornstein 2003; Barrie-Pittman 2004; Dubinsky 2007;

Rodrigues 2008; Boeckx, Hornstein, and Nunes 2010; Landau 2016a, b), we are still in a fix inasmuch as it is even difficult to reach a unanimous agreement as to what PC is, not to mention the fact that empirical data also vary. It seems (and I stress here the fragility of this claim) that PC is weirdly eclectic: characterizable by pragmatics, semantics and syntax although strictly semantic (cf. Culicover and Jackendoff 2005; Hornstein 2003; Pearson 2015) or pragmatic (cf. Bowers 2008) accounts have also been suggested.

At this juncture, it behooves me to introduce PC, the exemplification of which can be seen below:

(1) John1 told Mary2 that he1 wants [PRO1+ to meet in the morning].²

1 State Higher Vocational School in Włocławek, Poland, [email protected].

2 The linguistic context in (1) provides the necessary participants of the meeting other than the matrix controller. However, one should bear in mind that when such a context is missing, pragmatics will allow us to set up another one easily.

The relation between the non-lexical subject, standardly marked as PRO in syntactic theories, of the bracketed clause and the lexical subject of the matrix clause, i.e. John is not one of identity. The referent of the silent PRO seems to include John along with other individuals salient in the context (in this case it is Mary). Hence the notation 1+ on PRO.

According to Landau, only when the complement contains such collective predicates as meet, gather, congregate, assemble, the adverb together, can this effect be detectable. The requirements of semantics of collective predicates (their subject must be semantically plural but need not evince syntactic plurality) make us conclude that PRO must necessarily be PRO1+, the matrix nominal being incapable of functioning as the sole participant of the meeting. Since the matrix subject only partially controls the reference of the lower subject, this phenomenon has come to be known as partial control (PC), as opposed to Exhaustive Control (henceforth EC), where one can observe a strict referential identity between PRO and its controller:

(2) John1 managed [PRO1 to sell his house].

Since Landau’s (2000) seminal work marked a quantum leap in the study of this atypical species of control and hence most of the current analyses draw heavily upon the empirical observations described in the work, let me refer to his account when further characterizing PC.

At the heart of Landau’s analysis lies the claim that cross-linguistically PC is sanctioned only by factive, propositional, desiderative and interrogative verbs, while implicative, aspectual and modal verbs prohibit it, legalizing only EC. The following non- exhaustive list of control verbs is based on Landau (2000: 38):

(3) a. implicatives: dare, manage, bother, remember, forget, fail, etc.

b. aspectual: begin, start, continue, finish, stop, resume, etc.

c. modal: have, need, may, must, should, etc.

d. factives: regret, like, dislike, hate, loath, glad, sorry, etc.

e. propositional: believe, think, claim, deny, affirm, etc.

f. desideratives: want, prefer, promise, hope, yearn, agree, plan, insist on, refuse, etc.

g. interrogatives: wonder, ask, inquire, contemplate, interrogate, understand, know, etc.

And a sample of each verb class with PC is provided below:

(4) a. John1 preferred [PRO1+ to meet at six]. desiderative b. John1 wondered [where PRO1+ to meet at six]. interrogative c. John1 denied [PRO1 having met at six]. propositional d. John1 regretted [PRO1+ having met at six]. factive e. *John1 managed [PRO1+ to meet at six]. implicative f. *John1 began [PRO1+ to meet at six]. aspectual g. *John1 should [PRO1+ meet at six]. modal

Landau states that some contextual salience of the group about to meet is necessary to render (4a-d) licit but no context can rescue (4e–g).

All in all, PRO in this vision is semantically plural and it is essential that it co-occur with collective predicates. At first blush, this is a very neat explanation of what the phenomenon in question consists in. However, on closer inspection, it turns out that it is not devoid of maladies.³

This short paper is supposed to shake the analysis’s foundations by simply questioning the credibility of Landau’s main claim, i.e. PC PRO is licensed only by factive, propositional, interrogative and desiderative verbs and only when the complement clause contains a collective predicate. I will show that this thesis is of dubious standing since in Polish implicatives, modals and aspectuals sanction PC as well. In light of these facts I will argue that the semantic borderline established by Landau is erroneous and is not a decisive factor determining the availability of the ‘group’ interpretation of the non-lexical subject. I will also suggest that only inherently reciprocal verbs, and not collective verbs, can participate in PC. These verbs, taking a null discontinuous phrase, introduce an additional participant of the act denoted by the verb; hence the alleged ‘PRO1+ effect’.

On the whole, I will show that PC, the way we are used to perceiving it, does not exist. It is simply a result of the ellipsis of an argument of the reciprocal verb in a discontinuous phrase. Any attempts to make the phenomenon a semanto-syntactic regularity will produce a redundant segment of grammar, the existence of which is not vindicated. Finally, the data presented in the paper clearly call for a greater examination of more languages or even more thorough analysis of English.

2. SURPRISING EMPIRICAL OBSERVATIONS FROM POLISH

Polish non-finite clauses may be introduced by the element żeby ‘so that’ (5a) or they may occur without it, as seen in (5b):

(5) a. Marek chciał, żeby kupić mu samochód.

Mark wanted so-that to-buy him car

‘Mark wanted somebody to buy him a car.’

b. Ala lubi pracować w ogrodzie.

Alice likes to-work in garden

‘Alice likes working in the garden.’

However, in this paper I will confine my attention to the description of typical complementizer-free non-finite clauses containing either an infinitval or gerundive form.

Just like in English, in Polish we can also distinguish seven classes of different (in terms of their semantics) control verbs which take non-finite complementation (after Bondaruk 2004: 143–144):

3 The present paper focuses only on the central claim of Landau’s analysis, i.e. PRO in PC is semantically plural and it accompanies collective predicates. The syntactic part of the explanation, which is a natural by-product of the main assertion, is disregarded here. For a meticulous critical analysis of it see Snarska (2009).

(6) a. modals: musieć ‘must’, umieć ‘can’, powinno się ‘should’, mieć ‘be to’;

b. aspectuals: zaczynać ‘start’, kończyć ‘finish’, przestać ‘stop’;

c. implicatives: ośmielać się ‘dare’, zdołać ‘manage’, zapominać ‘forget’, pamiętać ‘remember’;

d. factives: lubić ‘like’, nienawidzieć ‘hate’, nie znosić ‘can’t stand’, być przykro ‘be sorry’;

e. propositional: powiedzieć ‘say’;

f. desideratives: chcieć ‘want’, woleć ‘prefer’, mieć nadzieję ‘hope’, obawiać się ‘be afraid’, zgodzić się ‘agree’, proponować ‘propose’, planować ‘plan’, decydować ‘decide’, zamierzać/mieć zamiar ‘intend’, myśleć ‘intend’, pragnąć ‘desire’, być skłonnym ‘be inclined’, być chętnym ‘be willing’;

g. interrogatives: zastanawiać się ‘wonder’, pytać ‘ask’, dowiadywać się

‘find out’, wypytywać ‘inquire’, domyślić się ‘guess’, zrozumieć

‘understand’, wiedzieć ‘know’, być jasne ‘be clear’.

In a manner similar to English, they can license either EC or PC. The former, with a full referential control of the matrix argument over PRO, is illustrated in (7):

(7) Janek1 zapomniał [PRO1 kupić kwiaty].

John forgot to-buy flowers

‘John forgot to buy flowers.’

As regards PC, it can be accessed in Polish with desideratives (8a), interrogatives (8b) and factives (8c), just in line with Landau’s theory:

(8) a. Ala zaproponowała Janowi1 spotkanie o szóstej, ale on1 chciał [PRO1+

Alice suggested John meeting at six but he wanted się spotkać o trzeciej].

REFL to-meet at three

‘Alice suggested to John a meeting at 6 but he wanted to meet at 3.’

b. Maria zadzwoniła do Marka1 z propozycją spotkania i Marek1 Mary called to Mark with suggestion meeting and Mark cały dzień się zastanawiał, kiedy [PRO1+ się spotkać].

all day REFL wondered when REFL to-meet

‘Mary called Mark with a suggestion of a meeting and Mark was wondering all day long when to meet.’

c. Maria1 zaproponowała Janowi2 spotkanie w “Bohemie”, ale on1 Mary suggested John meeting in Bohemia but he nie znosi/nienawidzi [PRO1+ spotykać

not stands/hates to-meet się w tej kawiarni.

REFL in this cafeteria

‘Mary suggested to John a meeting in the “Bohemia” but he can’t stand/hates meeting in this cafeteria.’

With respect to propositional verbs, in Polish, unlike in English, they take finite complementation only:

(9) a. Maria twierdzi, że pocałowała wczoraj Janka.

Mary claims that she-kissed yesterday John

‘Mary claims to have kissed John yesterday.’

b. Maria zaprzecza, że poszła z Jankiem do kina.

Mary denies that she-went with John to cinema

‘Mary denies going with John to the cinema.’

According to Bondaruk (2004), the only propositional predicate that allows a non- finite complement is powiedzieć ‘say’:

(10) Jan powiedział, żeby kupić mu samochód.

John said so-that to-buy him car

‘John said that he wants for somebody to buy him a car.’

However, in my opinion, this can’t be a propositional verb since, following Pesetsky (1982: 143), propositional verbs in general are diagnosed by the possibility of predicating truth/falsity of their complements:

(11) John claimed to have smoked pot yesterday, which was true/false.

By applying this test to (10), what we obtain is a semantic anomaly:

(12) #Jan powiedział, żeby kupić mu samochód, co było prawdą/fałszem.

John said so-that to-buy him car which was truth/falsehood

‘John said that he wanted somebody to buy him a car, which was true/false.’

The example above makes it clear that powiedzieć ‘say’, when followed by a non- finite complement, cannot be deemed propositional. What we get in (10) is powiedzieć

‘say’ in its another guise – that of a desiderative verb. Therefore, I will not consider powiedzieć ‘say’ to be a propositional verb (when it takes a non-finite complement) and assume, pace Bondaruk (2004), that PC is not licensed by Polish propositional verbs as these predicates simply do not subcategorize for non-finite domains.

So far there is nothing striking about PC in Polish. However, what does leave one astounded is the fact that Polish implicative (13a–c), aspectual (13d–f) and modal verbs (13g–h) allow PC in their complements:

(13) a. Jankowi1 udało [się PRO1+ spotkać o szóstej].

John-DAT managed REFL to-meet at six ‘John1 managed PRO1+ to meet at six.’

b. Maria1 tyle razy przypominała Jankowi2, że są umówieni Mary so-many times reminded John that they-are arranged na randkę w sobotę, ale on

for date on Saturday but he

mimo to i tak zapomniał [PRO1+ się spotkać].

despite this and so forgot REFL to-meet

‘Mary had repeatedly reminded John about their Saturday date, but he forgot about the meeting anyway.’

c. Maria1 była piękna i wyraźnie dawała mu2 do zrozumienia, Mary was beautiful and clearly she-gave him to understanding że ma ochotę na randkę,

that she-has willingness for date

ale Jan2 i tak nie odważył [PRO2+ się spotkać].

but John and so not dared REFL to-meet

‘Mary was beautiful and strongly implied that she felt like having a date but John did not dare to meet anyway.’

d. Janek1 miał problem ze znalezieniem odpowiedniego miejsca John had problem with finding right place na sekretne spotkania ale w końcu zdecydował, że

for secret meetings but in end he-decided that zacznie [PRO1+ spotykać się w tej starej szopie].

he-will-begin to-meet REFL in this old barn

‘John had a problem with finding a right place for the meetings, but in the end he decided that he would start meeting in this old barn.’

e. Janek1 i Marysia2 byli bardzo szczęśliwi podczas schadzek , ale on1 John and Mary were very happy during trysts but he niespodziewanie zaprzestał [PRO1+ spotykania się].

unexpectedly stopped meeting ^REFL

‘John and Mary were very happy during the trysts, but he unexpectedly stopped meeting.’

f. Jan1 nie znosi aranżowanych spotkań z dziewczynami. Wczoraj powiedział John not stands arranged meetings with girls yesterday he-told mi, że definitywnie kończy ze [PRO1+ spotykaniem się w ciemno].

me that definitely he-finishes with meeting REFL in blindly ‘John hates arranged meetings with girls. Yesterday he told me that he was definitely done with blind dates.’

g. Janek1 powinien był [PRO1+ spotkać się w eleganckiej restauracji, a nie w tej John should-have to-meet REFL in elegant restaurant and not in this

starej szopie].

old barn

‘John1 should have PRO1+ met in an elegant restaurant and not in this old barn’.

h. „Szef1 może [się PRO1+ spotkać o szóstej, ale nie o piątej]”, odpowiedziała Boss can REFL to-meet at six but not at five she-answered na moje pytanie o możliwość widzenia się z szefem.

for my question about possibility seeing REFL with boss

‘“The boss can meet at six but not at five”, she answered my question about the possibility of seeing the boss’.

The examples above clearly undermine Landau’s contention that implicative, aspectual and modal verbs are incompatible with PC. Needless to say, Polish points to a

different conclusion and as such it shows that, given a proper context, it is possible to generate the ‘group’ interpretation with ostensibly PC-resistant verbs. This, in turn, entails that PC is very much pragmatics-oriented and hence hard to control by syntax or semantics.

Polish is not the only hard nut to crack for Landau inasmuch as English also seems on closer observation more complex in this respect. However, since this contention is based only on the results of an informal survey that I carried out among ten native speakers of English, I leave this issue for further investigation.

There is also one more facet of the phenomenon under scrutiny that I would like to pay attention to as this is my firm belief that it matters greatly in proper understanding of PC and somehow it has been neglected in the debate.

3. COLLECTIVE PREDICATES VS INHERENTLY RECIPROCAL VERBS

Recall that Landau (2000) argues that PC is available only when the non-finite clausal complement contains collective predicates, i.e. meet, congregate, assemble, gather, etc. In particular meet is repeatedly evoked and used by Landau in his examples of PC.

However, the problem with this verb is that it is not a collective predicate in the sentences he provides.

The collective predicate is defined as a predicate that applies to a plurality of things as a whole and is never true of the individual members of the group (cf. Link 2008, Dowty 1987). Put differently, the so called “distributive reading” is impossible with collective predicates. Good exemplifications of such verbs are congregate, assemble or gather co- occurring only with semantically plural subjects, witness (14):

(14) a. They congregated/assembled/gathered in the morning.

b. The family congregated/gathered/assembled in church.

c. *John assembled/gathered/congregated in church.

(14a) shows that collective predicates are compatible with semantically and syntactically plural subjects, in (14b), although the subject is syntactically singular, its semantic plurality justifies the presence of the collective predicate and in (14c) the unacceptability of the sentence stems from the fact that the subject is both syntactically and semantically singular.

Now let us check how meet behaves when accompanied by different, in terms of their number, subjects.

(15) a. They met in the morning.

b. The family met in the morning.

c. *John met in church.

Admittedly, the verb follows the same pattern as the one presented in (14) and this warrants the conclusion that in this context it may represent a collective predicate.

However, it differs from the predicates in (14) in that it does not require a semantically plural subject (although, as seen in (15), it may take one). In such case it is forced to take

‘with DP’ phrase which is called a discontinuous phrase (see Dimitriadis 2008, Siloni 2012):

(16) John met with Mary in the morning.

The same cannot happen with truly collective predicates:

(17) *John congregated/assembled/gathered/ with Mary.⁴

This enigmatic state of affairs is accounted for if we characterize meet in (16) not as a collective predicate but as an inherently reciprocal verb which involves two participants/arguments playing the same role in the event.⁵ Thus the relation expressed between them is symmetric (cf. Siloni 2012), which makes it possible to change their syntactic position without causing a change in the sentence meaning:

(18) a. John met with Mary in the morning.

b. Mary met with John in the morning.

Summing it up, meet, congregate, assemble and gather cannot be perceived as representatives of the same species. Since congregate, assemble and gather invariably demand a semantically plural subject, they can be classified as truly collective predicates.

Meet, however, is an inherently reciprocal verb which may take a discontinuous phrase introducing an additional participant of the symmetric relation expressed by the verb.

The same reasoning with respect to collective predicates and inherently reciprocal verbs can be applied to Polish. Spotykać się ‘meet’ is an inherently reciprocal verb which may take on an intransitive shape, as in (19a) or it may be followed by a discontinuous phrase, as in (19b):

(19) a. Janek i Maria spotkali się.⁶ John and Mary met REFL

‘John and Mary met.’

4Admittedly, these predicates can appear with ‘with DP’ phrase but then they necessarily take semantically plural subjects, as seen in (a). This inference is based on an extraction carried out on January 30^th, 2018 from the Corpus of Contemporary American English. In the study the sequence gather/gathered/gathers with occurred 244 times, the sequence congregate/congregated/congregates with occurred 10 times and the sequence assemble/assembled/assembles with occurred 96 times. With none of these predicates was a semantically singular subject used. Importantly, the ‘with DP’ phrase in such constructions is a comitative and not discontinuous phrase. The difference between the two is that the former is an adjunct, whereas the latter is an argument (see Dimitriadis 2004, Siloni 2012).

This entails that the comitative phrase can be dropped. And it is in (b) below:

(a) They congregated/assembled/gathered with Mary in this old church.

(b) They congregated/assembled/gathered in this old church.

5 Other inherently reciprocal verbs include, for instance, kiss, hug, correspond or argue.

However, kiss and hug disallow the presence of discontinuous phrases:

(a) *John hugged with Mary.

(b) *John kissed with Mary.

Needless to say, they express reciprocity in intransitive constructions:

(c) John and Mary hugged/kissed.

6 In Polish, apart from spotykać się ‘meet’, reciprocity is expressed by such verbs as kłócić się

‘to argue’, korespondować ‘to correspond’, całować się ‘to kiss’ or przytulać się ‘to hug’.

b. Janek spotkał się z Marysią.

John met REFL with Mary

‘John met with Mary.’

Since the two nouns are in a symmetric relation, their order in the sentence can be reversed (compare (20) with (19b):

(20) Marysia spotkała się z Jankiem.

Mary met REFL with John

‘Mary met with John.’

As regards collective predicates, congregate, assemble and gather correspond to Polish gromadzić się i zbierać się. These verbs are compatible only with semantically plural subjects:

(21) a. Oni zgromadzili się/zebrali się w kościele.

they gathered REFL/assembled REFL in church ‘They gathered/assembled in church.’

b. Rodzina zgromadziła się/zebrała się po południu.

family gathered REFL/assembled REFL after noon

‘The family gathered/assembled in the afternoon.’

c. *Janek zgromadził się/zebrał się w kościele.

John gathered REFL/congregated REFL in church ‘John gathered/congregated in church.’

Unsurprisingly, they do not tolerate discontinuous phrases introducing an argument by means of the prepositional phrase,

(22) *Janek zgromadził się/zebrał się z Marią w kościele.

John gathered ^REFL/congregated ^REFL with Mary in church

‘John gathered/congregated with Mary in church.’

although allow comitative phrases, where ‘with DP’ phrase functions as an adjunct:

(23) Dzieci zgromadziły się/zebrały się z rodzicami/Jankiem w kościele.

children gathered REFL/congregated REFL with parents/John in church

‘The children gathered/congregated with their parents/John in church.’

On the whole, Polish and English are alike in that they both clearly distinguish between truly collective predicates and reciprocal verbs. To our needs it is important that both languages deem meet/spotykać się a reciprocal verb.

4. THE (OSTENSIBLE) PC – A NEW PERSPECTIVE

Throughout the paper I have been trying to show that the way PC tends to be described is not a proper way of approaching the phenomenon. First, PC is not licensed only by factive, propositional, interrogative and desiderative verbs. Second, meet and congregate, assemble,

gather cannot be lumped together and treated in the same way. They represent different types of verbs and hence should be addressed separately. Any account of PC that aspires to describe (let alone explicate) the phenomenon cannot overlook these two issues.

Needless to say, PC is a ‘recalcitrant control beast’, invariably hard to control and kudos to anyone who can offer a credible account of what we are dealing with. The solution I want to put forward is not a fully worked out proposal. It is a direction, an avenue that, in my opinion, is worth exploring. And even if it gets jettisoned by the linguistic world, it will still broaden our (control) horizons inasmuch as the critique obviously delivers much food for thought.

A starting point for my analysis is the claim that PC as a new exotic subtype of control does not exist. If a lexical element controls PRO, it means that there is a full referential dependence of PRO on its controller. So what we experience in a typical sentence with a ‘group’ interpretation is this:

(24) a. John1 wants [PRO1 to meet in the morning]

b. Janek1 zamierza [PRO1 spotkać się w tej starej stodole]

John intends to-meet REFL in this old barn

‘John intends to meet in this old barn.’

Notice that there is no ‘1+’ marking next to PRO, which means that PRO is both semantically and syntactically singular, just like its controller. But how to account for the

“group’ construal? It stems from the fact that meet/spotkać się, being a reciprocal verb, takes a discontinuous phrase in such constructions as (24). Recall that this phrase provides one of the two arguments/participants of the event denoted by the verb. Importantly, the phrase in our sentences with ‘group’ interpretation is silent, i.e., it undergoes ellipsis, which is illustrated in (25):

(25) a. John1 wants [PRO1 to meet with somebody in the morning].

b. Janek1 zamierza [PRO1 spotkać się z kimś w tej starej stodole].

John intends to-meet REFL with somebody in this old barn

‘John intends to meet in this old barn.’

And this entails the ‘1+effect’ in the non-finite subject position.

Meet is not the only reciprocal verb that can generate the meaning under consideration. The remaining reciprocal verbs such as argue, correspond, kiss or hug and in Polish kłócić się ‘to argue’, całowac się ‘to kiss’ or korespondować ‘to correspond’ can also give rise to ‘group’ interpretations and then the reasoning behind this construal is the same as in (25):

(26) a. Janek wiedział, że Maria1 chciała [PRO1 całować się z nim].

John knew that Mary wanted to-kiss REFL with him

‘John felt that Mary wanted to kiss.’

b. Janek czuł, że Maria1 chciała [PRO1 się kłócić z nim].

John felt that Mary wanted REFL to-argue with him

‘John felt that Mary wanted to argue.’

c. John1 was pissed at Mary and he1 wanted [PRO1 to argue with her].

d. John1 knows that he will miss Mary a lot and he1 is planning [PRO1 to correspond with her as often as possible].⁷

If reciprocal verbs are capable of generating ‘group’ meaning due to the presence of the null discontinuous phrase, this entails that collective predicates fail to license such an interpretation since they are incompatible with discontinuous phrases. And this prediction is borne out in (27):

(27) a. *John1 wanted [PRO1 to assemble/congregate/gather at six.]

b. *Janek1 chciał [się PRO1 zgromadzić/zebrać w kościele.]

John wanted ^REFL to-congregate/assemble/gather in church

‘John wanted to congregate/assemble/gather in church.’

5. WRAPPING UP

The so called PC is not a straightforward case. For still unknown, shrouded in mystery reasons some contexts allow it, whereas others disqualify it. But it is my hope that I managed to shed some light on the phenomenon by at least calling into question the validity of the central, widely accepted assumptions concerning the so called PC. I was trying to show that the empirical data from Polish invalidate the claim that the so called PC is licensed only by factive, propositional, interrogative and desiderative verbs. Furthermore, I drew attention to the fact that the proper characterization of this species of control will not do if we don’t discern between reciprocal verbs and collective predicates. Finally, I outlined my own perception of the phenomenon under consideration. More specifically, I suggested that only reciprocal verbs can license it by taking a null discontinuous phrase which introduces a silent participant of the act denoted by the embedded verb. All in all, whether my venture proved successful or not is not my job to assess.

Needless to say, this is by no means the end of the discussion of the so called PC.

There are still some loose ends and issues that I barely touched upon. For example, I did not discuss a responsibility relationship (Farkas 1988) between the matrix controller and the embedded action that may also have an impact on how we interpret non-lexical subjects denoting a group. I postpone digging into this matter and others for another occasion.

REFERENCES

Barrie, M., C. M. Pittman., 2004, “Partial control and the movement towards movement”, Toronto Working Papers in Linguistics, 22, 75–92.

Boeckx, C., N. Hornstein, J. Nunes, 2010, Control as movement, Cambridge, Cambridge University Press.

Bondaruk, A., 2004, PRO and control in English, Irish and Polish: A minimalist analysis, Lublin, Wydawnictwo KUL.

7As far as kiss and hug are concerned, recall that they disallow discontinuous phrases. I leave this issue unresolved for now as it requires much contemplation but I will come back to it in future research.

Bowers, J.,. 2008, “On reducing obligatory control to movement”, Syntax, 11, 125–143.

Culicover, P. W., R. Jackendoff, 2005, Simpler syntax, Oxford, Oxford University Press.

Dimitriadis, A., 2004, Discontinuous reciprocals, [Unpublished ms., Utrecht Institute of Linguistics, Utrecht, The Nettherlands.]

Dimitriadis, A., 2008, “Irreducible symmetry in reciprocal constructions”, in: E. König, V. Gast (eds), Reciprocals and reflexives: Cross-linguistic and theoretical explorations, Berlin, De Gruyter, 375–410.

Dowty, D. R., 1987, “Collective predicates, distributive predicates and ‘all’”, Proceedings of the 1986 Eastern States Conference on Linguistics (ESCOL), Columbus, Ohio State University, 97-116.

Dubinsky, S., 2007, On the syntax of exhaustive control and the calculus of events, Unpublished ms., University of South California.

Farkas, D. F., 1988, “On obligatory control”, Linguistics and Philosophy, 11, 27–58.

Hornstein, N., 2003, “On control”, in Hendrick, R., (ed.), Minimalist syntax, Oxford, Blackwell, 6-81.

Landau, I., 2000, Elements of control. Structure and meaning in infinitival constructions, Kluwer, Dordrecht.

Link, G., 2008., “The logical analysis of plurals and mass terms: A lattice-theoretical approach”, in:

R. Bäuerle, C. Schwarze, A. von Stechow (eds), Meaning, use and interpretation of language, Berlin, De Gruyter, 302–323.

Martin, R. A., 1996, A minimalist theory of PRO and control, Unpublished Ph.D. dissertation, University of Connecticut.

Pearson, H., 2015, “The semantics of partial control”, Natural Language and Linguistic Theory, 34, 691–738.

Pesetsky, D., 1982, Paths and categories, Ph.D. dissertation, MIT, Cambridge, Massachusetts.

Rodrigues, C., 2008, “Agreement and flotation in partial and inverse partial control configurations”, in: W.D. Davies, S. Dubinsky (eds), New horizons in the analysis of control and raising, Netherlands, Springer, 213–230.

Siloni, T., 2012, “Reciprocal verbs and symmetry”, Natural Language and Linguistic Theory, 30, 261–320.

Snarska, A., 2009, On certain issues of control in English and Polish: Partial control, split control and Super-Equi, Unpublished Ph.D. dissertation, Adam Mickiewicz University, Poznań.