1This presentation concerns those English multiword verbal constructions that come under various designations, namely COMPOUND VERBS, TWO-WORD VERBS, and, more often these days, PHRASAL VERBS. I shall call them only PHRASAL VERBS, leaving the other two designations for such compounds as short-change or manhandle, for example.
2The problem with phrasal verbs lies in their second element which is, for reasons that I do not find very clear, most of the times called a PARTICLE. According to some, a particle can be either a preposition or an adverb. If we believe others, it can only be an adverb (The verb+ preposition compounds are then simply called prepositional verbs).
3Some linguists establish a difference between ADVERBIAL PARTICLES and PREPOSITIONAL ADVERBS (Quirk et al, Cowie & Mackin, etc.). Bolinger even adds a fourth category, which he calls ADPREPS, like UP in he ran up (the pole) the flag. For Bolinger (1971: 28, note 5)
a PREPOSITIONAL ADVERB is a particle that can be either a preposition OR an adverb, whereas an ADPREP is a prepositional adverb which is a preposition AND an adverb at one and the same time.
4This type of classification in arrays of different categories and subcategories of particles seems to have emerged mainly because of a singular feature of these combinations : some are separable, and others are not. Although it is absolutely possible to say They pulled up the flag AND/OR they pulled the flag up, the same does not apply to They ran up a huge bill: the form *they ran a huge bill up is not a correct English utterance. Furthermore, if one is ready to accept, as is often the case, that RUN UP in he ran up a huge hill, also constitutes a phrasal verb, one also has to declare it non-separable, since the form *He ran a huge hill up is impossible. It is worth noting, however, that RUN UP, in this case (he ran up a huge hill), is technically nothing more than a plain prepositional construction.
5The issue of separability is generally considered to be the real heart of the phrasal verb matter, and to constitute the main stumbling block to any clear and simple explanation of the phenomenon. There are two kinds of explanations of separability in the literature: one is provided by lexicographers and the other by linguists. The first, and most widespread, notably in dictionaries of phrasal verbs, is, in my opinion, much less satisfactory than the second.
6These can be found essentially in dictionaries of phrasal verbs, and rely mostly on the assessment of the degree of idiomaticity of the verb + “particle” combinations. Their justification is summarized by Cowie & Mackin (1975: viii) as follows:
Discussions of idiomaticity are sometimes confused by introducing inappropriate grammatical criteria where considerations of meaning carry particular weight...
7My opinion is that considerations of meaning always carry the “heaviest” weight, but that the weight can only be carried appropriately only if a given speaker uses correct grammar to construct his message. Failing that, meaning is carried nowhere, and any attempt at communication is bound to founder.
8If we consider the four utterances below ‑ which I have borrowed from the Dictionary of Current Idiomatic English, vol. 1 Phrasal verbs ‑ and their classification in degrees of idiomaticity (see Table 1).
1. The machine turns on a central pivot
2. Our conversation turned on what was to be done when the battle was over
3. The caretaker turned on the hall lights
4. Pop music turns on many young people
9I find no reason to claim, as the authors do, that (2), for instance, is more idiomatic than (1), or (4) is more idiomatic than (3). The four utterances fit equally the commonly accepted definition of the word “idiomatic” (SOED 1965: 952):
Peculiar to or characteristic of a particular language; vernacular; colloquial.
10But their grammar does carry a significant weight, as acknowledged by the authors themselves (ibid: viii) :
... ON can be said to function as a preposition in the first example (the machine turns on a central pivot) and as a particle in the [fourth] (pop music turns on many young people)...
11My opinion is that ON, in pop music turns on many young people, functions as an adverb, not a particle. If one refers to SOED again, one soon notices that ON, in the example quoted, corresponds exactly, in terms of semantics, to the definition of the adverb (SOED 1965: 28):
Gram. One of the parts of speech; a word to express the attribute of an attribute; one that qualifies an adjective, verb or other adverb.
12It is true that this also corresponds to the definition of the particle, with no reference to meaning however, but only to size and position, when not used as an affix (SOED 1965: 1438):
Gram. A minor part of speech, especially one that is short and indeclinable, a relation-word; also a prefix or suffix having a distinct meaning, as -un, -ly, -ness.
13I hold that when meaning is at stake, calling an adverb or a preposition a particle only adds to the confusion, as may be exemplified by the heading GO TO in Courtney’s Longman Dictionary of Phrasal Verbs:1 42 different entries are listed under the main heading, none of which constitutes a phrasal verb. In all of them, the sole role of the “particle” is to link verb and complement. It does not add any semantic value to the verb itself in order to form a “phrasal verb” with it (here again I am referring to SOED for the definition of the word “phrase” (1965: 1492):
A small group of words expressing a single notion, or entering with some degree of unity in the structure of a sentence; an expression; esp. a characteristic or idiomatic expression.
14I shall mention only three examples in this paper (pp. 260-61), but all the entries listed have the same grammatical structure (verb + preposition + complement):
Go to the block: to be killed with an axe, as punishment
Go to the country: to hold a general election.
Go to law: to take someone to court about a disagreement, etc.
15In all these combinations, the added semantic value, if any, derives from metaphoric or metonymic interpretation of the complement, not from the preposition. At best, they can be described as “verbal phrases”, but certainly not as “phrasal verbs” like e.g., PUT UP in Yes, they could put up an itinerant poet for a few days (W. Boyd A Good Man in Africa, 1981).
16For the linguist, these are more satisfactory in that they do not rely on criteria of idiomaticity (which can only be subjective), or mere syntactic analysis, but on discourse analysis : the meaning of a given phrasal verb derives not only from the sum of the meanings of its components, but also from the circumstances surrounding its utterance.
17In this paper I shall restrict my analysis to the works of three authors: Dwight Bolinger, whom I need not introduce; Patrick Getliffe, who is the author of a thesis on English “verbs with particles”; Nigel Quayle, also the author of different works on verbs and particles and the syntactic “problems” they cause. (Getliffe and Quayle each belong to one school of linguistics, the former is a disciple of the French linguist Henri Adamczewski and his “meta-operational grammar”, and the latter, a disciple of Gustave Guillaume and his theory of the “psychomechanics of language”.)
18On the question of separability, Bolinger remarks that the rule of separability when the complement is a pronoun does not always apply (1971: 39), for instance in:
If you want to ease your mind by blowing up somebody, come out into the court and blow up ’ me.
I knew that the school board contemplated throwing out Spanish in order to throw out ’ me.
And he develops the following argument (Ibid: 41):
What needs to be asked is what it is that end position confers, and what it is about personal pronouns that makes them substantially less likely than nouns to have that something conferred on them. It is obvious from the examples that the pronouns are all contrastive.
- 2 The word “accent” is used here in the definition of, among others, O’Connor & Arnold (1963: 271): “ (...)
19The need to add contrastive value would thus result in “unusual” placement of the complement at the end of the utterance. One can bear in mind, however, that the speaker could very well have kept the complement at his “usual” syntactic place, and resorted to an unusual accent 2 to enhance the contrastive effect of the pronoun, as in:
I knew that the school board contemplated throwing out Spanish in order to throw ’ me out.
20a variant which is consistent with what he notes on utterances where the complement is a noun (Ibid: 55) :
If the noun retains the accent but the particle is put after it, the sensation most often seems to be just that of what is familiar under the circumstances.
21In Bolinger’s opinion, Where’s Jack? He’s taking his ’sister out tonight is a normal utturance, whereasWhere’s Jack? He’s taking out his ’sister tonight is less usual, the reason being that the complement his sister refers to somebody familiar, i.e. to an element of information already known (the addressee knows that Jack has a sister). On the other hand, he considers He’s taking out some friends tonight as a perfectly normal answer, because some friends refers to information unknown to the addressee.
- 3 The concepts of “news value” and “newsworthiness” were originally described by Peter Erades (1961: (...)
22This type of reasoning is much more convincing for the linguist than mere reference to unmeasurable degrees of idiomaticity. For Bolinger, an element can be considered “familiar” to a speaker when it bears no news value, or newsworthiness to him.3
23Quayle follows in Bolinger’s footsteps and expands on the idea of “news value”, which he defines as very close to the concepts of given information and new information developed by Halliday and Hasan (1976).
- 4 As far as I know no word has been coined in English to mark the difference between what French ling (...)
- 5 “Plus un élément nominal perd de sa valeur notionnelle, plus grande devient la possibilité - pour (...)
24Quayle is concerned with degrees of determination, not degrees of idiomaticity. For him, when a source4 considers that a given element of information has no “news value” or “newsworthiness” for the target, the degree of determination applied to that element is . In the case of phrasal verbs, he holds that the less notional value (therefore, the less degree of determination) the nominal complement of a given utterance has, the higher are the chances that the source will place it between the verb and the “particle” (1992-93: 62)5.
25He studies, for instance, the following utterance:
Tory urges M.P.’s to pull their sartorial socks up,
of which he says:
- 6 “L’intérêt de ce titre de journal réside dans le jeu de mots basé sur l’expression idiomatique pull (...)
This newspaper headline is interesting because it relies on a pun based on the idiomatic expression pull one’s socks up. In this case the noun socks does not actually refer to any article of clothing, it is a metonymic use of the word. The phrase was pronounced by a Conservative M.P. as an appeal to his colleagues whom he found badly dressed. Yet, any attempt at putting the noun-phrase sartorial socks back in its normal place ‑ the canonical position after the verb ‑ would result in an utterance that would be more difficult to interpret: Tory urges M.P.’s to pull up their sartorial socks.6
26He therefore thinks, like Bolinger, that separability does not depend on the complex verb itself but on the discourse value of the complement.
27Getliffe has a different analysis of the phenomenon. His conclusion, however, is very similar to Quayle’s. He argues that the association of a verb and a particle can result in what he calls, after Guimier (1980), a “verbe de langue”, in other words the lexicalisation in language of a semantic unit (verb+“particle”) used in discourse. Of these “verbes de langue”, he says (1990: 108):
- 7 “une fois constitué comme unité sémantique (pourra) devenir objet de discours, sa caractéristique é (...)
Once formed as a semantic unit [the “verbe de langue”] can become an object of discourse, whose particle can then be separated...7
28This, in my opinion, remains to be proved. If it were true, we would then have to accept that RUN UP in he ran up a huge bill is indeed an object of discourse, but an object of a different kind, whose particle is not separable.
29Whether separability or non-separability are explained in relation to the semantic value of the complement (Quayle), or as the consequence of an intrinsic characteristic of the complex verb itself (Getliffe) changes nothing to the facts : some verbs are separable and others are not. It is quite significant that Quayle’s conclusion should be very similar to Getliffe’s (1992-93: 66):
- 8 “La structure binaire du verbe à particule se révèle donc d’une importance considérable sur le plan (...)
The binary structure of verb+particle combinations proves to be of vital importance for the value of a given utterance : it gives the source a number of options not available to him if he resorts to a simple one-word verb, notably the possibility to place one element of the combination ‑ the particle ‑ in an unusual position in English, i.e. after the direct object.8
30As one can see, the question is no longer that of a complement being moved to the left, but of a particle that is moved to the right. Furthermore, one can dispute the statement that such a position of the particle (after the complement) is “unusual” in English.
- 9 In the Guillaume terminology: it is not “dematerialized”.
31Quayle’s analysis is valuable in that it explains why a phrase like he ran up a huge bill cannot be changed into *he ran a huge bill up, because the complement bill retains all its “notional” value.9 Yet, in the example he uses (Tory urges M.P.’s to pull their sartorial socks up), the word sartorial “rematerializes” the socks, so to speak. It gives its notional value back to the word socks (the source actually finds his targets badly dressed). And if the rule that he enunciates were applied strictly, the utterance should take the form Tory urges M.P.’s to pull up their sartorial socks, with no separation of the particle. If one looks up this particular “idiomatic" phrase in a dictionary, the form will almost always be pull one’s socks up, with a separated particle, but this is not true of all “idiomatic" phrases with pull up. Courtney (1983: 455) lists three in all, two of which have no separated particle:
Pull up one’s = Leave a place where one has lived for a long time.
Pull up stakes (AmE) = Leave a place where one has lived or worked.
32In these two phrases, the roots and stakes referred to are just as dematerialized as the socks of pull one’s socks up, yet the particle is not separated from the verb.
33If one looks up the same item in Cowie & Mackin’s dictionary (1975: 232), one finds:
pull one’s socks up = take command of oneself, become more purposeful and alert
34with a separated particle. The entry, however, is exemplified by the following excerpt from The Dream of Peter Mann by Bernard Kops (Penguin, 1970):
Penny: Please Peter, pull yourself together.
Peter: Leave me alone.
Alex: Pull up your socks
without separation of the particle...
35I believe that such strict rules cannot apply in “real life” situations. As Foss & Hakes have noted (1978: p.17) :
A theory of linguistic competence states the rules that are tacitly known by the speakers of a language. It is this knowledge that permits each speaker to make judgements about whether or not utterances are grammatical.
36Speakers accept forms of speech which are known to them as possible, and tacitly distinguish them from those which are not. Spoken language must always come first.
37To this day, I am not convinced (although I may well be wrong) that native speakers find much difference between utterances like I took my jacket off and I took off my jacket, at least when the complements are not “dematerialized”, although there is a difference since the word order is different. In the case of phrasal verbs (or any other two-word combination, for that matter), common sense ‑ plain diachronic analysis ‑ forces one to admit that any two words must exist before they can ever be combined. If one admits that there is a separability issue, I suggest the question be reversed. I propose to consider that phrasal verbs evolve from an original separate stage, through a non-separate one, to eventually become unseparable combinations, not the other way round.
38My reasoning is based on two central questions:
391. How can two separate semantic units combine and form what is known as a phrasal verb (a two-word combination)?
402. Is it really necessary to name the second word in the combination a “particle" when its only function is to modify the meaning of the verb?
These two initial questions lead to a set of two complementary ones:
411. Is it really necessary to distinguish between so many categories of “particles” (adverbs, prepositions, prepositional adverbs, adpreps and adverbial prepositions)?
422. How can one account for VERB + ADJECTIVE, VERB + NOUN or VERB + VERB combinations ‑ in which the role of the second word is no different from that of the “particle” ‑ without “overstretching” the meaning of the word “particle”; in other words, is HOME a particle in I think it was rubbing his feet with my hands which truly brought home to me our desperate position?
43In an utterance like I took my jacket off, OFF can be regarded as a “bereaved" preposition which has lost its complement (my back). The loss is, in fact, the result of a voluntary and purposeful omission by the speaker, who considers the explicit mention of the proform too obvious, and therefore useless. Its consequence is a change in the grammatical status of the preposition, which becomes an adverb. The utterance can be compared to other adverbial utterances like He threw the letter away, for example.
44I theorize that the frequent occurrences of such forms of speech has led to the possibility for speakers to regularly associate in their minds the meanings of two distinct entities (verb and adverb in this case) and form new semantic units (take off remove, throw away discard, etc.). A similar phenomenon can be observed with second words which, originally, are neither prepositions nor adverbs, but which, as soon as the association materializes, function as adverbs (push open, drive home, let fall, let fly, etc.).
45In my opinion, the genesis of phrasal verbs can be summed up in eight points:
461. The necessary condition for a phrasal verb to emerge is the separate existence of its two constituting parts: the verb and the second term.
472. If the second term is a real preposition, the emergence of a phrasal verb is impossible.
483. In the case of combinations where the second term is not a preposition (i.e. is not used to introduce a complement), the second term can only have one role: add semantic value to the verb. It is therefore, strictly speaking, an adverb; and phrasal verbs must, in my opinion, be considered as “adverbed” verbs.
- 10 Quasi-adverbs are markedly different in my mind from Bolinger’s adpreps. For Bolinger, UP is an adp (...)
494. The association of the meanings of two terms in any adverbed verb is dynamic. It must be viewed as an evolution towards irreversible inseparability. I propose to name this process “bonding”. The dynamic bonding hypothesis allows for the synchronic occurrences of unbonded and bonded forms (I took my jacket off and I took off my jacket), and also for the hybrid status of some second terms, which, even though they have acquired adverbial characteristics, have not completely lost their prepositional status. I propose to name second terms of this type quasi-adverbs 10. Depending on speakers, quasi-adverbs can be understood and uttered as either adverbs OR prepositions because they are adverbs AND prepositions at one and the same time. Such is the case, for example, of OVER and THROUGH in the following two examples:
“Sure,” Owen said. “And I suppose his father ran over the dog” (John Irving, 1989, A Prayer for Owen Meany, CORGI, 170)
Most of these kids can add but they have serious trouble thinking through simple problems. (Newsweek June 17, 1991)
50The different stages of bonding of a given combination in the minds of speakers depend on how far the dynamic bonding process has evolved towards its completion: the link between the two terms tends to become permanent, but cannot be considered so until the stage of actual bonding has been reached. Consider, for example, the following example:
Before they left the hut Wyeth went over one last time the recognition signals that the reception committee would use. (Ted Allbeury, 1991, The Dangerous Edge, New English Library, 175)
51where the position of OVER before one last time is an undeniable sign of its semantic bonding to the verb GO.
52As Bolinger showed, bonding is also a result-aiming process (1971: 96):
After something is bleached white it is white, and after a person gets away, he is away. The notion of resultant condition is essential to phrasal verbs.
53Thus, the verb RUN can form an adverbed verb with the adverb UP, but certainly not with UPWARD, whose semantic value does not imply that any result has been reached.
54The majority of second terms are one- or two-syllable words (UP, OFF, DOWN, IN, ON OVER) whose status was originally prepositional. But they are not the only ones: second terms can also find their origin in adverbs, verbs, adjectives and nouns: OUT, AWAY, FALL, GO, FLY, BELIEVE, WASTE, DEAD, EMPTY, HAPPY, PROUD, SILLY and HOME are perfect illustrations of the category.
55In the absence of bonding, one can only describe the occurrences as VERB + ADVERB sequences. These sequences, however, must be considered as the potential building blocks of adverbed verbs. Bonding occurs, most often, through figurative and/or metaphorical use of a given sequence, as TAKE OFF and PUT UP, for instance, in the following examples:
He felt about on the bed for his matches. I took the box off the mantelpiece, lit the match, held the flame to his cigarette. (Barbara Vine, 1990, Gallowglass, Penguin, 38)
Take your clothes off. / I won’t. / All right. We’ll go back to the other room and I’ll have them taken off. (Dashiel Hammet, 1930, The Maltese Falcon, Vintage, 206)
He took off his shirt and trousers and put on the red overalls. (William Boyd, 1981, A Good Man in Africa, Penguin, 254)
The Prime Minister’s party takes off for Brussels this morning for the next round of talks.(Cowie & Mackin 1975: 326)
Bill took off Winston Churchill to perfection. (Cowie & Mackin 1975: 326)
Here’s what happened to him. Going to lunch he passed an office-building that was being put up - just the skeleton. A beam or something fell eight or ten stories down and smacked the sidewalk alongside him. (Dashiel Hammet, 1930, The Maltese Falcon, Vintage, 65)
When Jeremy said he’d like next week off Manuel didn’t put up any objections and seemed interested when he said he was going to New York. (Ruth Rendell, 1985, The New Girlfriend & other Stories, Penguin, 85)
She and her boyfriend came and picked me up. And put me up for a few days. (Philip Roth, 1990, Deception, Penguin, 57)
56Bonding of a verb and an adverb is originally semantic, it then evolves through a phonic stage, and can eventually become graphic (notably in the case of nouns derived from adverbed verbs setup, makefast, hearsay, etc.).
57It is commonly accepted that, as regards pronunciation, phrasal verbs differ from plain verb+preposition sequences, in the accent that falls on their “particle”. In “Trees outside managed to stir in the moist tropical air”, IN is a preposition and is not accented, whereas, its adverbial counterpart is in “Please fill in the forms I have given to you”.
58There are, however, a number of circumstances when this distinction cannot be established: either because the preposition is monosyllabic and has no weak form : in “With Prairie hanging off him like a monkey in a tree...”, the preposition OFF carries an accent, or because the second term of the combination is what I have defined as a quasi-adverb, and its interpretation can vary with speakers. IN, OVER and ABOUT, for instance, can belong to this category, e.g.,:
We joined in the celebrations, ... you gloss over all the punch lines..., I set about Hereford as I had set about London.
59In the case of quasi-adverbs, speakers often resort to phonic means other than accent to show that bonding has materialized. They can mark a pause between the second term and the complement to show that, for them, the combination of verb and second term form a “sense group”, as defined by O’Connor & Arnold (1961: 272):
[A sense group is] a close-knit group of words preceded and followed by a pause but said without an intervening pause.
60Adverbed verbs are telling illustrations of the “sense group” phenomenon. In the case of:
“That’s just dishonesty,” said David, “that’s all that is. You mean that if you’re playing Tennessee Williams in Cheltenham you gloss over all the punch lines, for fear of offending the old ladies.” (Margaret Drabble, 1964, The Garrick Year, Penguin, 57)
61for example, some speakers show that GLOSS and OVER form a sense group, by “saying it without an intervening pause”, and marking a pause after the second term, effectively separating it from the complement which, as a preposition, it would have introduced (Busuttil 1994A).
62My purpose here is not, as one may well conceive, to have the elegant and widely-accepted “phrasal verbs” replaced by the far more awkward “adverbed verbs”. That would indeed be a Quixotic enterprise, especially on the part of a non-native speaker of English, unable, by definition, to grasp all the mystery and flexibility of the system. However, precisely because I am a non-native speaker, I too often find it very difficult to tread my way through the many misleading definitions available in the literature. Why not establish clear-cut basic distinctions between the different categories, do away with the word “particle”, and accept to call “adverbs” the second terms of adjectival, nominal or verbal origin when they function as adverbs? The basic classification I propose is as in Table 2.
STIR, as in Trees outside managed to stir in the moist tropical air,
PUT UP, as in Yes, they could put up an itinerant poet for a few days,
LET FALL, as in It was Fiona who, accidentally, let fall one evening the information to my parents that I had gone to evening school,
GO TO THE COUNTRY, as in If the members of Parliament vote against the government, they will have to go to the country.
63These three basic types of structures can be used by speakers practically ad libitum, since they are open to literal, figurative, metaphoric and metonymic usage.
64If the classification I propose were to be accepted, it would certainly not enable non-native speakers to always produce correct phrasal verbs of their own, but it would, no doubt, help them understand and master those they are bound to come across.