The hell in english grammar

The hell in English grammar
Isabelle Haïk
Université de Caen-CRISCO-EA 4255
September 2011. To appear. Abstract The DP the hell contains an emotive word and has no descriptive content. It has to find a place in syntax, while exercising its meaning on the proper part of the utterance that it bears on. Treating hell as a predicate targeting a Cause argument, the cause responsible for its utterance, and formally representing in the structure elements that contribute to its interpretation, such as the utterance situation, the hearer and the speaker, may yield a derivation that accounts for its interpretation. As a result of its syntactic position, in one construction, in which the hell is the specifier of a selected predicate, it takes the utterance situation as its Cause argument and in the other construction, which is a complex particle verb, it takes the event. The insertion of the hell in the syntax reveals a number of important facts, such as the difference between complex transitive structures and ECM-structures, it begs for the proper treatment of resultative constructions, and of fake object resultatives, and it points to the crucial relevance of selection for a predicate. Moreover, it requires questioning the argumental status of emotionally loaded DPs (scare the living daylights out of), with the role of the use of slang an important element in that question. Le DP the hell est formé sur une exclamation et n’a pas de contenu descriptif. La syntaxe doit lui trouver une place de telle sorte que ce DP puisse exercer son sémantisme sur la partie de l’énoncé sur laquelle il porte. En traitant hell comme un prédicat prenant un argument Cause, qui nomme la cause de sa production, et en représentant dans la structure les éléments pertinents à son interprétation, comme la situation d’énonciation, l’allocutaire et le locuteur, il est possible d’obtenir une dérivation qui rende compte de son interprétation. Pour des raisons de syntaxe, dans une des deux constructions examinées, où the hell est le spécifieur d’un prédicat sélectionné, il prend la situation d’énonciation pour argument causal et dans l’autre construction, qui consiste en un verbe composé (verbe à particule), l’évévement. La question de l’insertion de the hell dans la syntaxe permet de mettre au jour un certain nombre de faits importants, comme la différence entre la construction transitive complexe et la construction à Marquage de Cas Exceptionnel, elle oblige à proposer le bon traitement des constructions résultatives et des constructions résultatives à complément non sémantique, et elle note la pertinence cruciale de la notion de sélection pour un prédicat. De plus, elle exige de questionner le statut de la référence des DP chargés émotionnellement (comme dans scare the living daylights out of), où l’usage de l’argot peut signaler une émotion, à la manière d’une exclamation. The hell in English grammar
Isabelle Haïk
Université de Caen-CRISCO-EA 4255

Introduction

The hell is an expression of emotion that does not participate in the descriptive content of its
clause, which raises the question of its integration in the syntax of its clause and the way its
interpretation is obtained in the compositional semantic computation of the utterance it
appears in1. We will focus our attention on two particular contexts, characterized by the fact
that the hell occurs before a predicate. In such sentences, the hell expresses that the speaker
feels the emotion causing her/him to exclaim “hell!”, that emotion being caused either by the
intensity of the utterance situation, stressful or exciting, etc., or by the intensity of the event,
depending on whether hell takes as its argument the utterance situation or the event.
Construction A
Construction B

1. The data
1.1 Construction A. The hell occurs as a subject of a selected predicate
Semantically, construction A is interpreted with widest scope of the hell, and it reads as: ‘the
utterance situation is so intense that it causes me to utter “hell!”’. (In the whole article, when I
refer to the hell, I mean the hell and its class, for example, the shit, etc.) We will see that
construction A is possible only if the hell occurs before a predicate selected by a lexical (i.e.
not functional) head2. The verb can be transitive as in (6)-(9) or unergative as in (3)-(5). The
1 This article has been written in honour of Claude Guimier, with whom I am very glad to have collaborated. I have acquainted myself with pleasure with his general writing, and some less well-known, like his investigation on complex predicates, which finds echo in the present research. I have enjoyed his search for subtle and unexplored data and the interest he aroused for research at ELSAP and CRISCO and in our courses. Thanks to Ray Jackendoff for pointing out to me the work of Hoeksema and Napoli (2008). It is not before the completion of the article that this reference was brought to my attention They have been the first to propose to examine the two constructions together and under close scrutiny, so they are obviously granted primacy for this as well as their empirical coverage and their interesting discussion of taboo terms in general and the way they enter constructions. They put a crucial light on the history of the constructions. As for the theoretical analyses of the two constructions, our views and analyses have sometimes met but they more often diverge in the way we have expressed the phenomena. I have formulated the question of the form of these constructions in the minimalist framework, whereas they argue in favour of constructions as building-blocks, in which taboo terms like hell are used as intensifiers. 2 I leave for further research another context: (i) transitive verbs compatible with the hell are Quirk, Greenbaum, Leech and Svartvik (1985)’s class of complex transitive verbs, which select two phrases, a DO and a predicate bearing on that DO. The sentences below show that the hell (and the like) may insert before any class of predicate – an adverbial phrase or particle (get out of here, shut up in (3a) or (5)), an AP (keep quiet in (4)) or a PP (get in the house in (3a)) – but the predicate must always be selected by the verb, either closely, in the sense that it cannot be omitted, as with give up, leave AP, get PP, or more loosely, as when it is a locative adjunct in a construction with a verb implying a location like drift in (3d) or stay in (3b) – and whether the meaning is compositional (make worse in (9)) or not (give up in (5)). In this article, in long sentences, the hell and the like will be in boldface. I have left the errors because they can be revealing of the speaker’s care: (3) Get the hell out of here!/Get the hell in that house! Political parties and their organizations should stay the hell out of election
primaries. Let the people decide!! (internet) It is time to get the hell out of Iraq and Afghanistan! (internet) And if you had a brain in your head, you’d drift the fuck out of this town.
(Shoedog, George Pelecanos, Warner Books, p.140) Shut/give/grow/hurry/dress etc. the hell up!/Slow the hell down! etc. The Black Dahlias were sent to set the rock n' roll record straight. And set it the hell
why cant people just leave me and dillon alone and NOT MAKE IT THE
HELL WORSE!!! (internet)
The hell may not occur alone, whether the verb is intransitive or transitive: (10) It may not occur before an argument, whether or not introduced by a preposition: (11) *Give that money the hell to your sister! It may not occur before a non selected phrase: (12) *Do it the hell now!/while it’s time!/before the trial! They cut my mom RIGHT the hell open. (internet) I'd seal mine tight the hell shut (internet) the gate is wide the hell open on the GOP side of the aisle for Election 2012 (internet). Big Business and the Mysterious Machine is a band straight the hell outta Niagara-on- Such constructions are ruled out without the adjective before the predicate, which is the reason for my description in terms of selection of a predicate, because these structures are heads, the adjectives wide, etc., selecting a predicate, open, etc. Compare (i) with (12a). It does not seem to be compatible with unaccusatives (the imperative makes the verb agentive hence unergative, which is why some are not totally starred, but I will consider them bad): (13) *My ice-cream! It has melted the hell down! ?Sail the river before it freezes the hell up! Importantly, despite the presence of a selected predicate, the hell is impossible in Exceptional Case-marking (henceforth ECM) constructions: (14) *I believed/found/considered him the hell nice! And it is rejected in the other constructions built on selected predicates, namely, resultative constructions, predicative clauses, and subject- or object-secondary predicates3: Resultatives
*Don’t drink yourself the hell unconscious! *We wanted some rest and the clock ticked the baby the hell awake. Predicatives
*And I said to myself, appear the hell smart! Secondary predicates
In GB terms, predicative constructions are raising from a small clause (Couquaux 1979), resultative constructions are verbs complemented by a small clause (Hoekstra 1988) and secondary predicates are small clauses with a PRO subject (Stowell 1983), so they all indicate 3 Actually, some resultatives may enter construction A, such as kiss goodbye (Haïk 2011), though not for all speakers, and rip open, cut open, staple (and other verbs) shut, so more needs to be said about resultatives, though I will not solve this question here: (i) And I'm glad I can tell BOTH of us that 2010 is OVER and let's kiss it the hell
[…] rip this box the hell open!!! (internet) His friends should not let him speak, or they should, out of pity.staple his lips the
hell shut. (internet)
Then maybe you can look on the VERY FIRST PAGE OF THIS FORUM and see that your pipes are stuffed, locked, and welded the hell shut. (internet)
that the hell is incompatible with small clause constructions. We have reached the following conclusion, to be accounted for in two sections4: (18) Construction A. Descriptive generalization
a. The hell may only occur before a predicate selected by a lexical head. b. That predicate may not enter a small clause construction.
1.2 Construction B: the hell occurs before exhaustive out

(19)
In this construction, the hell has verbal scope: it is paired with the meaning of an intense process conveyed by the particle out, and it does not necessitate a stressful or exciting utterance situation as in construction A. Construction B reads as ‘the event I am describing is so intense that it causes me to utter “hell!”’, where the hell is speaker-related as concerns its utterance and the emotion felt. In this construction, the aspectual particle out always means exhaustion, a term I borrow from Chang (2004) for its accuracy. It means that the process goes as far as it can by exhaustion of some element involved in it. It does not always target the same argument, as we will see (sometimes the subject, sometimes the DO). (See also Lindner 1983, who explains out by the disappearance of something that used to be available to some perceiver). Out is not necessitated by the verb: But lots of verbs may optionally select it in order to express exhaustion of the process. This is shown below, the context being that the character is hitting at opponents: (22) I wanted to hit out until all I felt was numb. (Chasing the Dead, Tim Weaver, Kindle 4 Certain verbs select a manner adverbial, like behave or dress (Jackendoff 1972), so they are good candidates for the hell, but they do not license it, in fact: (i) *[…] mistook my intentions and snapped back at me, telling me to behave the hell
well and be nice. after that, i just decided to ignore her, . *I also dress the hell like a hipster, ride a ~sweet~ fixie and I don't own a Che shirt.
I have found those sentences on the internet but they have been judged starred. What is relevant is that speakers feel that the hell should be possible, presumably because the structure contains a selected predicate, but that predicate is not a complement of the verb but a verbal modifier, which must be what goes wrong here. Here are other attempts: (iii) *a mech as good as Bruno is entitled to behave the hell way he wants.
*She can dress the hell whichever way she wants.
*You should always dress the hell you want anyways.
Second, using Relational Grammar (Perlmutter 1980) terms, one striking property of this construction is the demotion-like behavior of the DO, which must occur as the object of out and not as a DO5: (23) Interestingly, fake-objects can also be “demoted” (see (58)-(59)). Consider: They yell the shit out of their lungs (internet) This made me cry the shit out of me (internet) They are body-part resultatives, the following, turned into construction B: (25) Not all body-part resultatives may give rise to construction B: (26) *He worked the hell out of his guts. (cf. He worked his guts out) *He argued the hell out of his heart. (cf. He argued his heart out) This construction is semantically constrained in such a way that it is best when the verb implies that some content goes off the body. Work and argue do not but yell and cry do, for yell, voice and air, for cry, tears. We will see in the discussion of (58)-(59) that this comes from the assumption that, in construction B based on such fake-object resultatives, out imposes a semantic restriction on its object. Other fake-object resultatives produce construction B. Consider the following (my butt and its class are kinds of reflexives, as shown in Collins, Simanique and Postal 2008): (27) I'll pull myself together and wait the hell out of my butt. (internet)
I pledge to sleep the hell out of myself tonight. Need to refill my damn snooze
I had a four-hour window of sleep last night which I was determined to sleep the hell out of. (internet)
I'm a bit tired but I hope the sleep the hell out of my trip tomorrow. (internet)
I ended up splurging should I work the hell out of myself or let it slide. (int.)
Wife and I just acquired a 2006 SeaSport Sportsman 22' that we plan to fish the hell out of for he next several years. (internet)
All these forms are fake-object resultatives whose object is realized as object of the particle: (28) a four-hour window of sleep which I was determined to sleep out. 5 This is the opposite of construction A, in which the DO remains a DO and cannot occur as object of the embedded predicate: (i) a SeaSport Sportsman 22' that we plan to fish out. Now, let us consider non transitive verbs. Construction B involves an intense event and not an intense utterance situation, so we have to search for that kind of meaning in the data. Unaccusative verbs like run out, dry out, etc. are ruled out (see (52) for an account): (29) *Let the fruits dry the hell out and then you can store them for months. Note that the agentive verb dry is fine, because it has the syntax of other transitive verbs, like beat, with out added to it6: (30) All proactive does for me is dry the HELL out of my acne and leave big red blotches
As for unergatives, they seem marginally possible. There are a few occurrences of the hell out strictly intransitive on the internet with intensity of the process rather than the utterance situation, though they are sometimes difficult to distinguish: to OP: when they roll it out, it's only for generic hand sets. operator branded will only get it after the operator.well.brands it (duh) so either you chill the hell out
and rebrand, or you just wait the hell out.
As for now, my ass is knocked outta energy and I just wanna go home and sleep the hell out.[…] PS: Did I say sleep the hell out? That should be "sleep the hell
In conclusion, we have established the following facts, to be accounted for: Construction B. Descriptive generalization
The hell only occurs before exhaustive out. The verb may be transitive or unergative but not unaccusative. When the verb is transitive, its DO, whether it is a true of a fake object, is realized as the object of out.
2. The analysis
2.1 Meaning and distribution of the hell
The DP the hell is formed with the exclamation “hell!”, so it carries some of its meaning (see
Wierzbicka 1992 and Ameka 1992a and b for semantic analyses of various kinds of
exclamations and interjections). The exclamation “hell!” is a linguistic reaction to something
that causes an emotion in the speaker, and its phatic function is to attract the hearer’s attention
to the cause of that emotion, whether explicitly or through implications. In the following
6 Agentive dry out, construed with the particle out, is fine in construction A too, meaning that in that case out is a selected predicate in a complex transitive structure, like transitive give up: (i) .cause you think since accutane dries out your skin it should dry the hell out a
dialogue, the sheriff signals to “T.T” something that provokes his emotion, a threat to police funding: (33) “Also, with this Moon Killer poop he’s taking a lot of focus off the school, which is where he doesn’t want the focus to be.” “You don’t live in New Lebanon, T.T. Hell, the school damn near pays our salary.”
(The Lesson of her Death, Jeffery Deaver, Bantam, 1993) There exists a metonymy which identifies the cause of a physical effect with that of the emotion causing that effect. For example, physical –ing adjectives like breath-taking characterize not a physical cause, but the psychological one which has the physical effect named by the adjective (Haïk 2005). The same happens here, exclaiming is a physical reaction provoked by an emotion. The cause of the emotion is the cause of the exclaiming. So, let us assume that in the DP the hell, the exclamation hell is a predicate that takes as argument the thing that causes its utterance. It is also evaluated with respect to the speaker and the hearer, so it is a three-place predicate, hell <C, S, H>, C the variable for the cause of its utterance, S for the Speaker and H for the hearer. Now, an interjection or an exclamation which is integrated in syntax has, at least, two ways to contribute its meaning to the sentence, or word, or phrase. One is delocution as with verbs like remercier ‘thank’, first analyzed by Benveniste (1966), in which the interjection converts into a verb having the metonymic meaning of doing what uttering the interjection performs (after first converting into a noun naming the act of uttering the interjection, in de Cornulier 1976). In Benveniste’s case, the interjection loses its expressivity, that is to say, it is not interpreted as uttered by the speaker. For instance, below, the speaker does not say “thanks!” nor “damn!” nor “goodbye!”: (34) Bob thanked Mary/didn’t give a damn/kissed his mother goodbye. The other case of delocution is the insertion in the clause of interjections or exclamations as uttered by the speaker, like fucking, crapping and others. In that case, the expression has the force of the exclamation. For example, these adjectives mean ‘causing me to utter the exclamatory word’, and the phrase that they modify is interpreted as the cause of the utterance of the exclamation (see Fradin 2003 for an analysis of French vachement ‘fucking’). As for the hell, it is of this kind. In construction A, the hell attaches to a selected predicate, and, by syntactic properties of post Spell-out internal merging, the utterance situation saturates its Cause position, yielding the reading of an intense utterance situation causing the utterance of the exclamation. The intensity of the utterance situation may iconically be indicated phonetically by emphatic stress or slow speed of delivery making each word, hence each meaning, salient. In construction B, the hell occurs within a complex verb whose e place saturates its Cause argument position, yielding the reading in which the process named by the verb, because of its intense quality, causes the utterance of the exclamation. Considering the syntactic position of the hell, a non referential DP with no descriptive content (Potts 2007), it merges in the only possible position for a non argument, a subject position, in the two constructions. 2.2 Construction A
2.2.1 Functional Pr

First of all, as a prerequisite to our account, I will take for granted the result of present-day
research that DPs are licensed in English only if they occur as specifiers of some functional
head before Spell-out, in order for that head to check the strong features of the DP, while at
the same time getting rid of its uninterpretable features by Agree. Here, evidence (sentences
like get her the hell out, with two DP objects) shows that the DP the hell does not have Case
features, but let us assume that it has other nominal features to check.
There are three situations to distinguish concerning the predicate fit for the hell: the predicate is selected, the predicate is not selected and the predicate is selected but in a small clause construction. The hell is possible only in the first situation, let us see why. I borrow Bowers’s (1993) Predicator, the functional head which takes predicates, further discussed in Bailyn (2002), though I do not borrow the whole theory, for instance, I will not use Pr to establish the subject-VP relation of the clause. I also assume that non selected predicates cannot be introduced by Pr and that Pr is optional in principle. So, an AP, etc., selected by a V can be found in [ V [ Pr AP ]PrP or it can be bare, [ V [ AP ] ], and a non selected predicate will necessarily be bare7. Feature checking of the spec by Pr does not necessitate any internal merging because the spec-head relation obtains when the two merge. As for small clauses, they are subject-predicate structures. Bowers (1997) shows that such structures can be obtained by the mere presence of Pr. I borrow his analysis, with some implementation: I assume that Pr has uninterpretable nominal features, requiring some DP to merge as its specifier, so that Pr checks the strong nominal features of the DP and gets rid of its own uninterpretable features. Thus, Pr must have a subject, small clauses are PrPs, and a predicate which selects a small clause selects Pr. All this is summarized in: (35) Only a predicate selected by a lexical head may optionally be introduced by Pr. Pr has uninterpretable nominal features. Let us consider the hell. This is a DP, so it must be the specifier of some functional head before Spell-Out. Given that it is not a referential argument, the most reasonable assumption is that it merges as the spec of Pr, a non theta position. Without Pr, the hell cannot have its features checked, which explains why it can only occur if the predicate is selected, given (35a). This happens in construction B too, where exhaustive out is a selected predicate, so it may be introduced by Pr, allowing the presence of the hell. Let us consider well-formed sentences with transitive verbs, like leave her the hell alone. Its V domain contains the argument-structure of the verb: arguments come from the numeration and saturate the argument positions of the verb in the stepwise fashion permitted by V-stacking (Larson 1988; Platzack 2001). As for the subject of the clause, it merges as the spec of the highest V: 7 In Chomsky (2007), a clausal subject is licensed by uninterpretable features of C inherited down by T, so here, selected predicates could differ from non selected ones by inheritance down by Pr of features originating in the selecting head. Inheritance by Pr would explain why selection has such a licensing effect on the presence of the hell while maintaining an identical external structure for all predicates. I haven’t had time to assess this analysis.
As for the possibility for the embedded AP to bear on the DO despite their distance, I assume
that the structure built by the hell is neutral for distance because the DP has no descriptive
content (an element n is neutral for some operation . if x . n = x), so the AP can recognize its
argument at the point when the DO merges in the structure.
We see in the tree above that, given that the only position compatible with the non referential status of the hell is the subject position of the embedded predicate, other word orders are ruled out: (37) The DO may not occur as complement of alone either, forming a construction B structure, because, among other things, construction B must be a resultative, and leave alone is not one: (38) *Leave the hell alone of her/alone out of her. Post Spell-out internal merging of the hell with intermediate specs allows it to saturate its argument positions along the way through adjunction structures. As an exclamative, by hypothesis, the hell takes widest scope. At LF, the DP symbolized by it naming the utterance situation in the spec of Sit(uation) saturates its C argument position, and the Speaker and Hearer functional categories have the S(peaker) and the H(earer) arguments as their specs, which saturate the argument positions of the hell when it adjoins to them, as illustrated below: As for intransitives, as in get the hell out of here, wake the hell up, etc. there is no DO. If the predicate is selected by V, it may be introduced by Pr. If it is, again the hell can be licensed by this functional head, which checks its nominal features when they merge: In the general case, when the hell is absent, as in wake (somebody) up, keep quiet, etc., the
structure will not be able to have optional Pr, as in [ wake [ up ]PartP ]V’, because Pr would
have uninterpretable features not eliminated by Agree with a DP in its spec.
2.2.2 Construction A: why the hell is not possible in small clause constructions

Let us recall from (18b) that the hell may not occur in small clause constructions:
(41)
*Why don’t you believe [ her the hell smart ], for once! *I should have been born [ the hell rich ]! Let us see why. An ECM verb has only one complement, PrP. Here is an ECM-structure: In such a structure, the DP her and the emotive DP the hell compete for the same spec position, X, making merging of the hell impossible. The difference with complex transitive structures is that, there, the DP is a true DO, with which the V can combine once it has combined with PrP, as in (36). In that case, the hell may merge as the spec of Pr. Whereas in ECM-constructions, the verb has one complement, PrP, and all the elements of the PrP must have merged when it merges with the verb. The external argument of the A merges as the spec of Pr, which leaves no room for the hell. Otherwise, supposing that, instead of the semantic subject of the AP, the hell were to merge as the spec of Pr, the argument of the adjective would have to be realized as its complement, as in construction B, but then this does not have the desired meaning of her interpreted as the subject of smart: (43) The same analysis holds for all structures involving a small clause, as in resultatives ((44a) below and the next section) or predicative sentences or sentences with secondary predicates ((44b) below). Secondary predicates require PRO as subject, competing with the hell: (44) One more note concerning the resultative. Given that, as we will see in the next section, reanalysis applies in resultative constructions, they threaten to allow the realization of the DO as the internal argument of the embedded predicate, yielding the following, but, again, if clean takes an internal argument, it should go with the meaning that whatever is clean has gotten rid of the table, which is not the meaning we suppose the sentence should have: (45) 2.3 Construction B
2.3.1 A lexical resultative with exhaustive out

Let us summarize the analysis of resultatives that we will use. Resultatives, like wipe clean,
are complex verbs, as discussed in Dowty (1979), created by a verb and a predicate heading a
small clause (Hoekstra 1988). I assume that thematic reanalysis, in which the theta-grid of the
complex verb results from the union of the grids of its heads as in Jayaseelan (1988), applies
between the V and the SC. Importantly, Hoekstra’s idea is that it is not the verb that is
responsible for the syntactic presence of its complement(s) but the SC. We will thus suppose
that in resultatives, the verb cannot stack up. As for the subject of the verb, it merges as spec
of V. Reanalysis yields argument-sharing, similarly to Baker (1988)’s and Baker and Stewart
(1999)’s analysis of serial verb constructions. For discussion on complex verbs, see for
example Winkler (1996) and Mc Intyre (2007). The present analysis of resultatives is close to
Bowers’s (1997), I fully agree with his concern for a syntax mirroring compositional
semantics, but I diverge slightly in having reanalysis and thus allowing argument sharing.
According to me, in wipe the table clean, the table is not only the syntactic and semantic
subject of the PrP but it also at the same time saturates the argument position of the verb.
Construction B is obtained by forming a resultative complex verb with a verb, the intermediate Pr head and the aspectual particle out. I will assume that this particle is only aspectual and does not have any theta-role to assign. Its theta grid is empty but it conveys the meaning of exhaustion. Union of the grids of the verb and the particle is such that the complex verb inherits the two arguments of the verb, an agent x and a patient y: scare has an agent and a patient, and out is empty, so the union of {agent, patient} and {} is {agent, patient}. First, because the particle is selected by the verb, it may be introduced by Pr, and because this is a resultative construction, by Hoekstra’s hypothesis, it must be (the predicate added to V must be PrP). Pr has nominal uninterpretable features which must Agree with some DP, forcing merging of a DP as its spec. The starting structure is thus the complex-verb form, where X merges from the numeration as spec of Pr: (46) Let us now consider what X can be. First, it can be the semantic DO of scare and we get usual resultatives, in which X saturates the y position of Pr. Second, X can be the DP the hell. Let us see what happens then. Semantic reanalysis between the verb and the particle results in a series of heads which are all bits of the same lexical element, the verb scare Pr out, all bearing the semantic information of the whole unit. Now, because the hell is licensed by Pr, and because Pr is an instance of the complex verb, this makes the hell indirectly licensed by the verb, which is close to a DO property. Using Chomsky’s claim that DOs must be theta related to their verbs, I will assume that, because the hell is licensed, even indirectly, by V, it must be in a semantic relation with it. Not having descriptive content, the hell cannot be an argument of either of the heads of the complex verb. Let us assume that it is semantically licensed through saturation of its C(ause) position by the e argument position of Pr inherited from the verb. That properly yields the reading in which the event e causes the utterance of the interjection, and that explains why it is only when the hell merges with a complex verb that it takes e as its argument: (47) Now, scare has an internal argument. Given the hypothesis that V-stacking is not allowed in resultative constructions, neither the V nor the DP may remerge up for saturation. In fact, that would produce the following ill-formed sentence, so we know it is not an option: (48) (=(25)) *He scared me the hell out. In resultatives in general, it is the embedded predicate which is responsible for the presence of the internal arguments of the verb. This happens here too. The aspectual particle out may structurally take an internal argument. By reanalysis, it has the same grid as the verb, so the relation between the particle and the DP can be licensed by saturation: (49) 2.3.2 The semantics of the verb in construction B

When out expresses exhaustion, we expect the internal argument to measure the event (Tenny
1994 and Dowty 1991), but in this construction, exhaustion of the event can be measured with
respect to all its participants, the subject or the DO, which is one of the reasons I suggested
that out has an empty theta grid, it does not seem to target any specific DP. Construction B is
preferred with verbs of physical or psychological impact, whose subject or object arguments
are humans from which something involved in the event escapes, like energy or serenity, etc.:
I bet he resented the hell out of all her charity work. (Deadman’s Switch, There is some creation with that construction, like the following examples found on the internet, though acceptability judgments qualify some of them as extremely marginal: (51) (*)I would buy the living hell out of this $62000 hermes motorcycle. Who can edit the living hell out of this picture? Google Promoting the Hell Out of YouTube A community about kirk franklin rocks the hell out of his version of september. “I’ll do that. Right after I call my lawyer and tell him to get ready to sue the hell out of the government.” (Dead Sleep, Greg Isles, Penguin (SIGNET), p. 308)
However, construction B has been the source of a few morphological forms, such as the
adjective (scared/bored) shitless and the compounds shit/hell-scared, shit-bored, shit-tired,
shit-beaten, etc., in which the base verbs are psych verbs or the verb beat. Construction B has
been created on a psych verb, scare, and it is very much used with the verb of physical impact
beat. If we suppose that morphology is more conservative than syntax, the restriction of these
morphological constructs to verbs of impact confirms the restriction of construction B initially
to such verbs.
2.3.3 Construction B: why it is impossible with unaccusatives
As described in (32), construction B rejects unaccusative verbs:
(52)
??The supplies have run the hell out (*of). However, construction B may passivize, yielding a structure similar to the unaccusative one: (53)
In GB terms, the trace of the unaccusative sentence would be Case-marked by out and not in
the passive one. Conditions on NP-traces require that they do not bear Case, excluding (52)
and ruling in (53). The same idea can be implemented in the minimalist framework though I
will not attempt to give a technical account here.
2.3.4 Construction B is a resultative structure

This construction may not only contain exclamations, like hell, (be)jesus, (living) Christ, but
also body-related nouns like wits and life, etc:
(54)
This scared the hell/shit/fuck/fucking hell/living hell/bejesus/life/wits/living shit/living daylights/piss/crap/living Christ/etc. out of me. I’d have beaten the pulp out of him. (The Turtle Warrior, Mary Relindes Ellis, Sentences with the hell seem to have appeared first. According to the Oxford English Dictionary, there is a written occurrence of get the hell out in 1911, one do I knock hell out of you (J. Joyce) in 1922, while the shit and others occurred later, for instance in 1966 an occurrence of let’s beat the shit out of him. We may assume that when the exclamation “shit!” entered the construction, it opened the way to body-related DPs, because this word can be interpreted either as the exclamation or the referential noun here, with the two readings, ‘this scared me to the point that it causes me to utter “shit!”’ or ‘this scared me to the point that the shit run out of me’, the descriptive reading. Consequently, the semantic frame of construction B allows body-related nouns in the position of the hell, yielding the meaning that one’s body is exhausted of its content under the effect of the intensity of the process named by the verb. However, importantly, I claim that the DP must remain an emotive word and cannot be a full-fledged argument. This is shown by the rejection of emotively neutral terms like tears, showing that referential DPs are excluded8: (55) This is correlated with the fact that the verbal expression is always a hyperbole, not a faithful description of reality: the shit, daylights, piss, etc. do not really run out of the body. The expressions of bodily contents found in this construction are slang words, for their emotive force, as we will see, like piss, crap, etc., or nouns naming the extreme limits of a body or of life, like wits, daylights and pulp, some of which appear with that meaning only in that construction, like daylights. So, I assume that, like the hell, they are not referential arguments and they are treated like emotive DPs (see the section on emotive –ing words below). These body-part DPs are impossible in construction A, confirming that they belong to construction B because of their lexical meaning, even though they are not treated grammatically like ordinary arguments: (56)
2.3.5 Construction B with a resultative as input

This section shows that construction B, quite productive with transitive verbs, can be formed
on resultatives, and even fake object resultatives.
Let us consider how resultatives with exhaustive out react with insertion of the hell. First, when transitives take semantic objects with exhaustive out, like dry DP out, insertion of the hell produces the same structure as transitive constructions like scare the hell out of DP, in which the DP contains something that escapes incrementally and exhaustively, as in (30), repeated here, so nothing more needs to be said: (57) All proactive does for me is dry the HELL out of my acne and leave big red blotches
Now, let us consider construction B formed on fake-object resultatives mentioned in 8 I leave open the question of why construction B is bad with non emotive words, as in (55). It looks like a Specified Subject condition violation because a referential DP subject of the particle prevents the relation between the V and the object position of the particle. In other words, realization of the object of the verb as object of the particle is possible only it this does not cross over a referential subject, but I leave the technical solution for further research. Fake-object resultatives are formed from merging a small clause PrP with a V. V and PrP reanalyze, meaning that their theta-grids unite. Given that the DO is not an argument of the verb, the presence of this DP must be due to the particle. So, we must assume that out in fake-object resultatives has a semantic argument, and that its theta-grid is not empty. I assume that this argument, y, may be either a container, like his eyes, or a content, like the trip. Reanalysis results in the argument of the particle becoming the argument of the new complex verb: Now, let us consider what happens if the hell merges as the spec of Pr and not his eyes or the trip. When the particle takes a complement, this complement may name a content, as in be out of cigarettes or a container, as in be out of the room. So, because they are compatible with the semantics of the particle, his eyes and the trip may merge as complements of the particle, granted that they name either a container out of which some content escapes (tears for the container eyes), or a content, for instance a trip whose effects escape from the mind and body. So, realizing the DP as the complement of the particle instead of its subject does not change the semantic relation between the particle and the DP, given the previous assumption that the syntactic subject of out may be a container or a content9. The structure is: 9 It could be that out is an unaccusative particle, similarly to Svenonius (2003), discussed in Levin and Sells (2007), and that the insertion of the hell simply prevents the DP from merging up, but I have not had time to investigate this solution. So, here I adopt an analysis in which exhaustive out of a complex verb is reversible, with DP out expressing the same thing as out of DP. Second, the data tells us that the hell takes scope over the event and not the utterance. As with true-object resultatives, that is because the hell is in a spec-head relation with Pr, and Pr is an instance of the V by reanalysis. This forces a semantic relation between the two, again saturation of the C argument position of the hell by the e position of the verb carried by Pr. To summarize, concerning the role of the particle out and the effect of the insertion of the hell, with true-object resultatives, the particle out assigns the theta role coming from the
verb by reanalysis, whereas in fake-object resultatives, the particle assigns its own theta role.
As for the effect of reanalysis, it makes the hell indirectly dependent on V, forcing Pr and the
hell to be in a semantic relation, which explains why hell takes e as its argument.
2.3.6 The necessary presence of the hell for the internal realization of the argument

In the previous sections, we saw that, when the hell is inserted, the internal DP argument may
be realized as the object of out. But we have not yet seen why this argument may not be the
object of out in the absence of the hell. That is to say, the internal realization of the DP occurs
if and only if the hell is present:
(62)
*This scared out of her (vs. this scared the hell out of her) *They beat out of him. (vs. They beat the hell out of him) *I’ll sleep out of my brains. (vs. I’ll sleep the hell out my brains) *He cried out of his eyes. (vs. He cried the hell out of his eyes)
Resultative constructions are by definition built with PrP, a small clause, and small clauses
require a specifier (cf. also the original claim of Simpson 1983 and Levin and Rappaport 1995
that resultatives always have a DO). Pr has uninterpretable features which can erase only
under Agree with its spec, so, it is not possible to overstep the need for Pr to have a spec,
which these sentences do.
3. Miscellaneous remarks
3.1 Syntactic properties of construction B
Passive of the hell or other nouns is possible, as well as passive of the complement:
(64)
We have Superheroes getting beaten the hell out of in this issue (internet) That's why I'm so good today--the hell was scared out of me at an early age. The poor girl just got the crap scared out of her (Tess Gerritsen, Keeping the Passive of the hell confirms that passive may apply to non arguments, and passive of the complement shows that the hell out of is not a determiner like a lot, as shown by the impossibility of passivizing from quantifiers, even if it seems to carry the meaning of a quantifier: (66)
3.2 The types of clauses in which the hell is inserted
Construction A is most often found in clauses naming injunctions, in direct speech, or indirect
speech with hell evaluated with respect to the participant’s situation:
Political parties and their organizations should stay the hell out of election
Because hell takes the utterance situation as its argument, it favors clauses that actively
involve the utterance situation. Conversely, construction B does not favor imperatives but
statements because emphasis on the predicate can better be put on sentences that describe
what is the case.
3.3 Emotive –ing words
As we saw, the interjection “fuck!” allows the formation of the adjective fucking. Let us take
for granted the morpho-semantic rule producing such a form10:
(70)
[X-ing]Adj means ‘causing me to utter “X!”’, in which X is an exclamation the utterance of which is the physical reaction to some emotion. This rule is responsible for adjectives like shitting and crapping, whose bases are the exclamations “shit!” and “crap!”. They occur in places where the hell would have been possible, that is to say, a wh context, as in (71) and construction B, as in (72): (71) What the shitting fuck? (internet) (72) I don't quite know what is going on in this picture but it's starting to scare the Just a few years ago, that very thought would scare the CRAPPING SH!T outta In these constructions, even though some nouns are not exclamations, they take –ing, like piss: (73) What in the pissing hell does this mean (internet) OK. NOW HOW THE PISSING HELL DO I GET DOWN? (internet: picture of the author on the flank of a vertical icy rock face) I want to see that captain Mayuri die. He annoys the living pissing hell out of I could only watch liek 40 seconds of this n he started to irritate the pissing hell I will also say the same for living, in the living shit/hell/etc, in which live is not the verb but the noun life, found in construction B (he scared the life out of me). That is obvious from the semantics, for in these occurrences living does not have a meaning related to the act of living. The word living may be an intensifier in construction B, as in (74). Interestingly, this word may also, for some speakers, be an intensifier in other constructions than construction B, which use the hell, as with wh in (75), even if wh does not accept the life, as in (76): 10 The two –ing forms effing and frigging are not created by this rule, for neither *f…! nor *frig! are exclamations. They are created on the word fucking, the euphemism effing on its form and the vulgar word frigging on its meaning. But how the living hell did you get that amount of coins? *But how the life did you get that amount of coins? One may wonder how all this is possible, first, how words like life and piss may act like exclamations and second, how living may, for some speakers, occur where life may not. I claim that it is the existence of construction B which makes this possible. Consider the exclamations and the nouns found in construction B11: (77) bejabbers)/life/wits/living shit/living daylights/piss/crap/crikey etc. out of me. Construction B is a frame in which the subject of the embedded predicate may be an exclamation but also a DP naming some content of the body. However, in that case, this noun must be heavily loaded emotionally (recall (55)). Vulgar or slang words like piss are sometimes used under strong emotion. Moreover, coprolalia, involuntary swearing, is an instance of it as a disorder, showing that there is a link between emotion and the utterance of slang or dirty words. We thus conclude that in construction B, the use of slang is caused by strong emotion. And the same happens with words naming heavily loaded notions like life and death, or sanity, or extreme ideas, such as pulp (cf. to the pulp) or daylights and wits, whose plural marker contributes to the meaning of intensity by suggesting that all the parts essential to life are concerned (Emmanuelle Roussel, p.c.). Like taboo words, they are emotionally loaded, and their use, in construction B, is provoked by strong emotion. Then, the process of delocution, in which an exclamation forms a word whose meaning involves the utterance of the exclamation, extends to slang words, like piss and emotionally loaded words, like life, wits, daylights, etc., which may be found in places normally filled with exclamations, as in construction B. So, first, rule (70) above should integrate the possibility for a noun that is not exclamatory to fill the role of X in X-ing forms: (79) X-ing means ‘causing me to utter X,’ in which X is an exclamation or an emotionally loaded word. Emotionally loaded words which can signal an emotion through their use are slang or vulgar words, or words having to do with taboo subjects, like sex, death or
This rule is responsible for the creation of living, pissing and others because life and piss are
emotionally loaded in construction B. Second, for some speakers, these words become
independent from the construction that has made their existence possible, and they occur
elsewhere with the same meaning, ‘causing me to utter this heavily loaded word’.
4. Conclusion

In conclusion, construction A (get the hell out, give the hell up, etc.) is formed with selected
predicates of all sorts, whether the verb is transitive or intransitive. I have assumed that,
optionally, Pr may take selected predicates and that the hell merges as the spec of Pr. Small
clause constructions, such as ECM constructions and others, may not host the hell because the
subject of the small clause and the hell compete for the same position.
11 There is some variation among speakers, some accept the living shit but not *the living hell. Construction B (scare the hell out of) is a resultative construction, in which out means exhaustion, and in which the heads V, Pr and out reanalyze as one semantic unit. The non referential DP the hell merges as the spec of Pr and, owing to reanalysis, the argument of the verb scare may be realized as the object of out. When the argument is not a semantic argument of the verb, as with fake objects resultatives, it is the semantic argument of the particle. DPs related to the body may, on the condition that their use reveals some emotion, function in the same way as the hell in construction B. The hell has the utterance situation as its Cause argument in construction A and e in construction B. In construction A, the interpretation of the hell is effected in the structure via
adjunction to specs that contain its arguments, which supposes that elements like the utterance
situation can be structurally represented as elements taken into account in the semantic
computation of an utterance.
REFERENCES
AMEKA F., « Interjections: The universal yet neglected part of speech », Journal of
Pragmatics 19(2), 1992, p. 101-108.
AMEKA F., « The meaning of phatic and conative interjections », Journal of Pragmatics 18,
1992, p. 245-271.
BAILYN J., « Overt predicators », Journal of Slavic Linguistics, 10/1-2, 2002, p. 23-52.
BAKER M., « Object sharing and projection in serial verb constructions », Linguistic Inquiry
20, 1989, p. 513-555.
BAKER M. and OSAMUYIMEN T. S., « On double-headedness and the anatomy of the clause »,
ms. Rutgers University, 1999, online at
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.39.6340&rep=rep1&type=pdf
Benveniste E., Problèmes de linguistique générale, Paris, Gallimard, 1966.
BOWERS J., « The Syntax of Predication », Linguistic Inquiry 24, 1993, p. 591-656.
BOWERS J., « A binary analysis of resultatives », Proceedings of the 1997 Texas Linguistic
Society Conference, Texas Linguistic Forum 38, 1997, p. 43-58.
CHANG C., « All syntaxed out: the exhaustive verb-particle construction in English »,
ms.University of Maryland, 2004, online at
http://terpconnect.umd.edu/~cbchang/papers/allsyntaxedout.pdf
CHOMSKY N., « Approaching UG from below », SAUERLAND U. and GÄRTNER H-M. (eds.),
Interfaces + Recursion = Language? Chomsky’s Minimalism and the View from Semantics,
Berlin, Mouton de Gruyter, 2007, p. 1–29.
COLLINS C, SIMANIQUE M. et POSTAL P. M., « An AAE Camouflage construction »,
Language 84, 2008, p. 29-68.
COUQUAUX D., « Sur la syntaxe des phrases prédicatives en français », Linguisticae
investigationes 3, 1979, p. 245-284.
DE CORNULIER B., « La notion de dérivation délocutive », Revue de linguistique romane 40,
1976, p. 116-144.
DOWTY D., Word Meaning in Montague Grammar. Dordrecht, Reidel, 1979.
DOWTY D., « Thematic Proto-Roles and Argument Selection », Language 67, 1991, p. 547-
619.
FRADIN B., « Delocutive adverbs in French », 10th International Morphology Meeting,
Szentendre, Magyarorszag, 2003.
HAÏK I., « Marrant et les adjectifs comme ça », Lingvisticae Investigationes, XXVIII-2, 2005.
HAÏK I., « Kissing goodbye », ms. Université de Caen, 2011, online at
http://www.crisco.unicaen.fr/IMG/pdf/Kissing_goodbye-dec-2009-1_interligne.pdf
HOEKSEMA J. and NAPOLI D. J., « Just for the hell of it: A comparison of two taboo-term
constructions », Journal of Linguistics, 44, 2008, p. 347-378.
HOEKSTRA T., « Small clause Results », Lingua 74, 1988, p. 101-139. JACKENDOFF R., Semantic Interpretation in Generative Grammar. Cambridge, MIT Press, 1972. JAYASEELAN K. A., « Complex Predicates and Theta-Theory », WILKINS W. (ed.). Syntax and Semantics 21, Semantic Relations, San Diego, California, Academic Press, 1988. LARSON R., « On the Double Object Construction », Linguistic Inquiry 19, 1988, p. 335-391. LEVIN B. and RAPPAPORT M., Unaccusativity: At the syntax-lexical semantics interface. Cambridge, MA, MIT Press, 1995. LEVIN B. AND SELLS P., « Unpredicated Particles », REDPILL, Hee Wee L. and Uyechi L. (eds.), Center for the Study of Language and Information, Stanford University, 2007. MCINTYRE A, « Particle verbs and argument structure », Language and Linguistics Compass,1(4), 2007, p. 350-367. OXFORD ENGLISH DICTIONARY, 2nd edition, prepared by Simpson J.A. and Weiner E.S.C, Oxford, Oxford University Press, 1989. PERLMUTTER D., « Relational Grammar », MORAVCSIK E. A. and WIRTH J. R. (eds.), Current Approaches to Syntax (Syntax and Semantics 13). New York, Academic Press, 1980, p.195-229. PLATZACK C., « Multiple Interfaces », NIKANNE U. and VAN DER ZEE E. (eds.), Cognitive Interfaces: constraints on linking cognitive information Oxford: Oxford University Press, 2001, p. 21-53. POTTS C., «The expressive dimension », Theoretical Linguistics 33, 2007, p.165–197. QUIRK R., GREENBAUM S., LEECH G. and SVARTVIK J., A Comprehensive Grammar of the English Language, New York, Longman, 1985. SIMPSON J., « Resultatives », LEVIN B., RAPPAPORT T. and ZAENEN L. (eds.) Papers in Lexical Functional Grammar, IULC, 1983. STOWELL T., « Subjects across categories », The Linguistic Review 2, 1983, p. 285–312. SVENONIUS P. « Limits on P: Filling in Holes vs. Falling in Holes », Nordlyd (Troms1 Working Papers in Linguistics) 31, 2003, p. 431-445. TENNY C., Aspectual Roles and the Syntax-Semantics Interface, Dordrecht, Kluwer, 1994. WIERZBICKA A., « The Semantics of Interjections », Journal of Pragmatics 18, North-Holland, 1992, p. 159-192. WINKLER S., Focus and secondary predication, Studies in Generative Grammar, n° 43, Berlin, New York, Mouton de Gruyter, 1996.

Source: http://www.crisco.unicaen.fr/IMG/pdf/The_hell_in_English_grammar-2.pdf