On Sun, 19 Oct 1997, Carolyn Kirkpatrick wrote:
>
>         We looked up the address
>            ~~~~~~~~~
>
> and adjectival subject complements, e.g.,
>
>         You look pretty.
>
> We are told that some of these complements can be prepositional
> phrases:, e.g.:
>
>         You look like hell.
>
> But then we are presented with this example:
>
>         Ryan looks like his older brother.
>
> We are told to analyze this example, too, as
> subject / verb / adjectival phrase.
>
> It doesn't "feel" the same to my students nor, I must confess, to
> me.  While "like hell" can be paraphrased as "awful," we cannot
> think of a suitable paraphrase for "like his older brother."
 
There's an easy test for deciding whether you've got a well-entrenched
phrasal verb vs. verb plus PP. Unfortunately, it isn't straightforward for
these cases, but here it is anyway, for other cases: Replace the object of
the prep. with a pronoun. The order of pronoun/prep tells you:
 
We looked up the address.
We looked it up. -- phrasal verb
*We looked up it.
 
We looked up the flue. -- verb + PP
We looked up it.
*We looked it up.
 
Also, try moving the PP to the front of the sentence:
*Up the address we looked. Doesn't work for phrasal verbs.
Up the flue we looked. Doesn't work for non-phrasal verbs.
 
Now:
Ryan looks like him.
*Ryan looks him like. suggests 'look like' is not a phrasal verb. But:
 
??Like him Ryan looks. Suggests that it is!
 
Why the mixed results?
 
Like so much else in our real language (and unlike so much in traditional
grammatical analysis), the status of phrasal verb is a matter of degree --
some verb - PP constructions are squarely phrasal verbs, and thus permit
pronoun + P; others are _on their way_ to becoming phrasal verbs and may
not yet permit the permutation. It is typical of in-between constructions
that they pass some of the tests but fail others, while prototypical
examples pass all tests. Constructions on their way to becoming phrasal
verbs have an idiomatic 'feel' to them. And, as Martha Kolln points out,
may be somewhat idiomatic but also somewhat predictable in their meaning.
Also, such in-between cases become the subject of conflicting analyses like
the case here. It is also possible that one person's grammar differs slightly
from another's, and some people might 'feel' strongly that we have a true
phrasal verb, while others feel we don't. Some might rightly feel that it
is in between.
 
I think there is a further complication in that 'look like hell' is a
conventional expression -- a collocation (an expression that is pretty
predictable in meaning as the sum of its parts, but nevertheless occurs
with such high frequency that it is virtually a lexical item in itself).
It probably is a better idea not to compare the 'look like' resemble with
this kind of example.
 >
> *We* think "looks like" here is one of those phrasal verbs.
> Certainly Ryan (NP1) and his older brother (NP2) are two different
> people.  And you could say
>
>         Ryan resembles his older brother.
>              ~~~~~~~~~
> Why isn't this a subject / verb / direct object pattern?
 
It _is_ a subject/ verb/ direct object pattern. Like status as a phrasal
verb, transitivity is also a matter of degree, and the degree of comfort a
native speaker feels with the passive version reflects just how transitive
the verb is. Try it this way:
 
Ryan is resembled by his older brother.
 
This sounds better to me than:
>
>         *His older brother is resembled by Ryan.
>
This is probably because of the 'his' in the first NP, which precedes its
co-referent NP. How about:
 
My father is resembled by many people.
 
Sounds very awkward to me, but lands between 'Ryan is ..' and 'His older
brother is ..'
 
Grammaticality judgments are a matter of degree and can also differ from
one individual to another (which makes intro syntax classes lots of fun!!)
 
The insistence in traditional grammar and in some modern linguistic
theories on all-or-nothing category membership just doesn't reflect the
reality of language. Every language is undergoing change, and many things
are on their way to becoming something else, but not quite there. Language
change takes generations. Also, meanings are gradable, like transitivity.
To 'see' someone is not as active as to 'hit' someone; to 'resemble'
someone is even less active. This means that grammatical analysis is not
always black and white, cut and dried. But if the reality isn't, why
should we expect our analysis to be?
 
If anyone is interested in technical reading sources on these topics, I
can supply some titles.
 
Please let me know if this explanation helps. Thanks!
 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Johanna Rubba   Assistant Professor, Linguistics              ~
English Department, California Polytechnic State University   ~
San Luis Obispo, CA 93407                                     ~
Tel. (805)-756-2184  E-mail: [log in to unmask]      ~
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~