Particle Retention in Deverbalized Constructions

Jane Rosenthal
UC Santa Cruz

My objective in this paper is to show that verb+particle combinations have the best chance at becoming a deverbalized construction (id est, the verb is used as a noun and/or adjective while retaining the particle); whereas constructions with verb+prepositions have no chance for grammaticality in deverbalized constructions. But what is the difference between a particle and a preposition and how can I tell what I am dealing with?

While prepositions and particles use the same lexical entries (e.g. in, at, up, around, etc), they differ in how they can be used and what processes they can undergo. In general terms, a preposition is more closely dependent on the noun phrase that follows it, whereas a particle acts almost like an adverb by modifying (and usually changing) the meaning of the verb. I have found several diagnostics to determine whether we are dealing with a particle or a preposition.

Diagnostic #1: The simplest diagnostic is to strand the P-word (I use the term P-word to refer collectively to prepositions and particles). To do this, we move the P-word to the position immediately after the noun phrase, or apply the passive (deleting the agent), or form a sentence without an NP object of the P-word. All of these methods work to strand the P-word. If we get a grammatical result after stranding the P-word we are dealing with a particle; if it is not grammatical, it's a preposition. Contrast the following sets of examples:

(1)a.  Alan Greenspan turned around the economy.
 b.  Alan Greenspan turned the economy around.
 c.  The economy turned around.
(2)a.  Steve Jobs makes around $20 million per year.
 b.* Steve Jobs makes $20 million per year around.
 c.* Steve Job's make around is $20 million.

Set (1) illustrates an instance of a particle. In (1b), I moved the P-word to the position after the noun phrase and yielded a grammatical result. Also, I formed a grammatical sentence without a NP to the P-word. Set (2) on the other hand, shows an instance of preposition use. Moving the P-word yielded an ungrammatical result, as did applying the passive.

In some cases it is difficult to determine the status of the P-word. Observe:

(3)a.  Cristina looked up the address.
 b.  Cristina looked the address up.
 c.  The address was looked up.

At first up appears to be a particle, however,

 d.  The naughty boy looked up Cristina's dress.
 e.* The naughty boy looked Cristina's dress up.
 f.# Cristina's dress was looked up.

I will explore what is going on here in the next section of this paper; however, I feel that this situation shows that there is a need for other tests to determine the status of the P-word.

Diagnostic #2 involves wh-movement. Pied-piping (and by this I mean the P-word moving with the noun phrase to the spec of C position ) results in an ungrammatical result when we have a particle. The result of this movement shows the dependent relationship between the preposition and NP; and demonstrates the dependent relationship between the verb and particle. Contrast the following set of examples:

(4)a.  Joe relies on his girlfriend.
 b.  On his girlfriend Joe relies.
 c.  On who(m) does Joe rely?
(5)a.  Alfonzo the clown blew up balloons.
 b.* Up balloons Alfonzo blew.
 c.# Up what did Alfonzo blow?

The example set in (4) show use of a preposition as shown by the grammaticality of pied-piped phrases involving wh-movement. The example set in (5) is more problematic. Pied-piping is not allowed in (5b), but with a wh-word, we get an odd result. While (5c) seems grammatical, the answer to the question cannot be balloons which was the target answer. Instead, the only possible answer would be a pipe or some other object like that, so the meaning is not the same as what we intended.

A third and less definitive diagnostic is related to the first diagnostic that I presented in which we move the noun phrase to the position before the P-word. In this diagnostic, if we replace the noun phrase with a pronoun and try to move the pronoun to the position before the P-word, we have a particle if we get a grammatical result. We get an ungrammatical result if we are putting it before a preposition.

(6)a.  Emily looked after the children.
 b.* Emily looked them after.
(7)a.  The student finished up his homework.
 b.  The student finished it up.

I stated that this diagnostic is less definitive than the other tests because not all particles can allow the particle to be followed by a pronoun. For example:

(8)a.  Joe screwed it up.
 b.* Joe screwed up it.

But the pattern of (8) passes all of our other diagnostics for including a particle, e.g.

 c.  Joe screwed up the phone lines.
 d.  Joe screwed the phone lines up.
 e.  The phone lines were screwed up.

So this diagnostic is rather limited in how it can be applied. It can merely help to verify that it is a particle, but cannot be used conclusively. It is useful however, for the following example:

(9)a. The failing student fucked up the exam.
 b.The student fucked the exam up.
 c.The exam was fucked up.
 d.The student fucked it up.

Here, fuck up passed all of our diagnostics, and so we must be dealing with a particle. Question: And what is the point of this?

My hypothesis states that only particles can be used in deverbalized constructions. What I mean by this is that if a verb+particle combination is deverbalized (used in a form that acts as a noun or adjective), we have a good chance of getting a grammatical result. Conversely, my hypothesis predicts ungrammaticality of deverbalized constructions of verb+prepositions in this type of construction. I will use the examples given above to illustrate this.

(10)a.? The looked-up number was in the phone book.
 b.  Smith is a commonly looked-up name.
 c.* The phone number was in the look-up.

Previously, I showed some difficulty in determining whether look up contains a particle or a preposition. My belief is that some P-words are pragmatically dependent on what kind of noun phrase they interact with and this determines what the result will be in a deverbalized construction. In this example, it is possible to use it as an adjective in a limited context (10b), but not in other contexts (10a) or as a noun (10c) showing that it is pragmatically controlled by the noun phrase that follows it.

(11)a.* The make-around salary for Steve Jobs is $20 million.
 b.* Steve's make-around is $20 million.
(12)a.* My relied-on mother arrived late.
 b.* My mother is my rely-on.

Sets (11) and (12) demonstrate two examples that I already determined to contain prepositions, and neither example works in deverbalized constructions as predicted by my hypothesis.

Sets (13)–(18) show constructions that pass the particle tests and have deverbalized constructions.

(13)a. The turn-around time was short.
 b. The economy's turn-around was dramatic.
(14)a. Isabelle felt turned-on by Aspects of the Theory of Syntax by Noam Chomsky (especially Chapter 2).
 b. Noam Chomsky was a turn-on.
(15)a. The student felt put-down by Suzanne's comments.
 b. Suzanne's put-downs hurt the student's feelings.
(16)a. The blow-up furniture stuck to my butt.
 b. George had a major blow-up at the party.
(17)a. The screwed-up wires caused an electrical fire.
 b. The fire was the result of a screw-up.
(18)a. The fucked-up exam confused the students.
 b. The student is a fuck-up.

The diagnostics that I presented work well for sets (13)–(18) and provide a good basis for the grammaticality of these examples and the ungrammaticality of sets (11) and (12).

A special case of my hypothesis is how deverbalized idiomatic expressions can be accounted for. One example that is somewhat problematic is come on. This combination fails all of the diagnostics:

(19)a.  Marguerite came on a boat.
 b.* Marguerite came a boat on.
 c.* Marguerite came it on.
 d.* A boat was come on.

And yet, we do have a deverbalized usage:

 e.  Austin Powers' line "Let's shag" was a come-on.

At first it seems like this should totally go against my hypothesis because it fails the diagnostics that I have laid out so far (namely the preposition cannot be stranded). This phrase seems very idiomatic, as the use of the following imperative:

 f.  Come on!

With this example, my theory seems applicable, since this shows that the P-word can be stranded, thus upholding the principle of the first diagnostic and shows that it does have a chance to be realized in a grammatical deverbalized construction.

Another example that has this same problem is drop out.

(20)a.  Lloyd dropped out of school.
 b.* Lloyd dropped school out.
 c.* School was dropped out.
 d.* Lloyd dropped it out.

But we do have a deverbalized construction:

 e.  Lloyd was a drop-out.

But, again, we do have an imperative form with a stranded P-word, as found by this legendary quote from Timothy Leary:

 f.  "If you take the game of life must turn on, tune in, and drop out."

And, so, again, the hypothesis continues to hold together if we expand the notion of stranding the particle to include imperatives.

In conclusion, I feel that there is strong evidence to give an effective accounting of why a construction such as turn around can be used as an adjective and a noun, but a construction like make around cannot. This accountability comes down to a simple test as to whether there is a particle or preposition in the verb phrase which would lead to a prediction of grammaticality in deverbalized constructions. Once the status of the P-word is established (if it is a particle or a preposition), we work under the hypothesis that preposition constructions are unable to be used for deverbalized constructions, but conversely, particle constructions have a good chance of resulting in a grammatical deverbalized construction.