Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Phrasal Verbs

82 views
Skip to first unread message

DKleinecke

unread,
Apr 7, 2013, 2:49:53 PM4/7/13
to
In the recent discussion of phrasal verbs nobody mentioned
http://www.usingenglish.com/reference/phrasal-verbs/
which is a list of 3274 English phrasal verbs. They list 214 of them
as ending in "on" including "sleep on" which they define as meaning
"think about something" as in "My boss said she'd have to SLEEP ON it
when I asked her for a raise." Each entry is marked as separable or
inseparable.

They define: "A phrasal verb consists of a verb and a preposition or
adverb that modifies or changes the meaning; 'give up' is a phrasal
verb that means 'stop doing' something, which is very different from
'give'. The word or words that modify a verb in this manner can also
go under the name particle."

Then they give as examples: "Phrasal verbs can be divided into groups:
Intransitive verbs - these don't take an object
They had an argument, but they've made up now.
Inseparable verbs - the object must come after the particle.
They are looking after their grandchildren.
Separable verbs - with some separable verbs, the object must come
between the verb and the particle:
The quality of their work sets them apart from their rivals."

Finally they add: "With some separable verbs, the object can before or
after the particle, though when a pronoun is used it comes before the
particle:
Turn the TV off.
Turn off the TV.
Turn it off."

This, I assume, can be considered mainline. Elsewhere they call these
idioms. I imagine they are using compositionality as the test. That
is, if the meaning of the verb plus the particle(s) is not the
combination of the two meanings then it is an idiom.

But compositionality is a blunt tool. Even the "meaning" of a verb is
elusive but the "meaning" on a particle is often very hard to pin
down. The problem is that English verbs seem best handled by
describing the structures
they appear in in terms of templates. For example (# here means a
phrase with an -ing form verbal)
see O
want O to #
etc.

This generalizes phrasality but makes almost all verbs have at least
one alternative usage with a non-empty template and therefore,
essentially, phrasal. It follows that I would drop "phrasal verb" as a
technical term and, instead, enter all the templates in the lexicon.

wugi

unread,
Apr 7, 2013, 4:21:46 PM4/7/13
to
Op 7/04/2013 20:49, DKleinecke schreef:
I would, on the contrary, restrict the notion of phrasal verbs to those
that take emphasis on the particle (ie not including those that do so on
the verb).
Correct me if I'm mistaken in assuming that in
- I have to slEEp on it,
- They are lOOking after their grandchildren,
the emphasis is on the verb.
In which case it's not phrasal verbs to my PTS (personal tongue sense:-).
Admittedly, this is an opinion of a non-E. mothertongue speaker, one who
is allowing ourselves to compare phrasality with the case of separable
verbs in German and Dutch, at that;-o

guido
http://home.scarlet.be/~pin12499




Glenn Knickerbocker

unread,
Apr 7, 2013, 4:25:57 PM4/7/13
to
On Sun, 7 Apr 2013 11:49:53 -0700 (PDT), DKleinecke wrote:
>Finally they add: "With some separable verbs, the object can before or
>after the particle, though when a pronoun is used it comes before the
>particle:
> Turn the TV off.
> Turn off the TV.
> Turn it off."

I would say that was true of (nearly?) all of them.

�R Around here, the fun is always filled with blanks.
http://users.bestweb.net/~notr/arkville.html --Theresa Willis

Brian M. Scott

unread,
Apr 7, 2013, 5:44:02 PM4/7/13
to
On Sun, 07 Apr 2013 22:21:46 +0200, wugi <br...@brol.be>
wrote in <news:kjskgo$jh7$1...@speranza.aioe.org> in
sci.lang,alt.english.usage,alt.usage.english:


> I would, on the contrary, restrict the notion of phrasal
> verbs to those that take emphasis on the particle (ie not
> including those that do so on the verb).

> Correct me if I'm mistaken in assuming that in

> - I have to slEEp on it,
> - They are lOOking after their grandchildren,

> the emphasis is on the verb.

You're right about the first one; for me the second has
approximately even stress.

[...]

Brian

DKleinecke

unread,
Apr 7, 2013, 7:13:43 PM4/7/13
to
On Apr 7, 2:44 pm, "Brian M. Scott" <b.sc...@csuohio.edu> wrote:
> On Sun, 07 Apr 2013 22:21:46 +0200, wugi <b...@brol.be>
The shifts in emphasis, of course, often occur. But the criterion for
an idiom to be a phrasal verb idiom (phrasal verb is an unfortunate
name because we are talking about verb particle pairs or triple) is
composition. There are many phrasals with two particles - three would
be possible but I know of no examples,

Nathan Sanders

unread,
Apr 7, 2013, 7:45:41 PM4/7/13
to
In article
<8cecc3e3-a1cf-47ea...@g4g2000yqd.googlegroups.com>,
DKleinecke <dklei...@gmail.com> wrote:

> But compositionality is a blunt tool. Even the "meaning" of a verb is
> elusive but the "meaning" on a particle is often very hard to pin
> down. The problem is that English verbs seem best handled by
> describing the structures
> they appear in in terms of templates.

It seems like you've rediscovered subcategorization, which linguists
have known about for quite a few decades.

Nathan

--
Department of Linguistics
Swarthmore College
http://sanders.phonologist.org/

Robert Bannister

unread,
Apr 7, 2013, 8:37:52 PM4/7/13
to
I confess I have only ever considered verbs to be "phrasal" if they are
(to use the terminology above) separable. That's why I mentioned in an
earlier post that "sleep on" fails the "it" test. I don't see why the
other type need a special name as they are merely verbs followed by a
preposition. It is true that sometimes the preposition may be regarded
as an adverb, but then that is the problem with English: the parts of
speech regularly defy attempts to put them in boxes.

--
Robert Bannister

Brian M. Scott

unread,
Apr 7, 2013, 9:08:40 PM4/7/13
to
On Sun, 7 Apr 2013 16:13:43 -0700 (PDT), DKleinecke
<dklei...@gmail.com> wrote in
<news:afda5a69-1475-4411...@q9g2000yqd.googlegroups.com>
in sci.lang,alt.english.usage,alt.usage.english:

> On Apr 7, 2:44 pm, "Brian M. Scott" <b.sc...@csuohio.edu> wrote:

>> On Sun, 07 Apr 2013 22:21:46 +0200, wugi <b...@brol.be>
>> wrote in <news:kjskgo$jh7$1...@speranza.aioe.org> in
>> sci.lang,alt.english.usage,alt.usage.english:

>>> I would, on the contrary, restrict the notion of phrasal
>>> verbs to those that take emphasis on the particle (ie not
>>> including those that do so on the verb).
>>> Correct me if I'm mistaken in assuming that in
>>> - I have to slEEp on it,
>>> - They are lOOking after their grandchildren,
>>> the emphasis is on the verb.

>> You're right about the first one; for me the second has
>> approximately even stress.

> The shifts in emphasis, of course, often occur. But the
> criterion for an idiom to be a phrasal verb idiom
> (phrasal verb is an unfortunate name because we are
> talking about verb particle pairs or triple) is
> composition. [...]

True, but wholly irrelevant: I was answering a question
about pronunciation, nothing more.

benl...@ihug.co.nz

unread,
Apr 7, 2013, 10:33:46 PM4/7/13
to
I would suggest, rather, that it's the adverbial type that is
separable (passes the 'it' test).
Adverbs also show a potential for stress, which prepositions (at least
the common monosyllabic ones) don't.
"Sleep on it" does not show separation because "on" is a preposition
here.
"Sleep on" (= continue to sleep) does not even qualify for the test,
since it can't take an object; but the "on" is clearly an adverb with
a consistent meaning in "walk on", "party on" etc.
"Putting me on" (perhaps obsolete now? but very common a generation
ago) would be an example of an adverbial transitive with "on".
I believe (checking Crystal's Dict of Ling & Phon) that the
conventional understanding of the term "phrasal verb" includes both
the prepositional and adverbial types. The key thing is that they are
idiomatic lexical units.

Robert Bannister

unread,
Apr 7, 2013, 11:02:35 PM4/7/13
to
Then I would call them "compound verbs" as in German.

--
Robert Bannister

DKleinecke

unread,
Apr 8, 2013, 9:04:24 PM4/8/13
to
On Apr 7, 6:08 pm, "Brian M. Scott" <b.sc...@csuohio.edu> wrote:
> On Sun, 7 Apr 2013 16:13:43 -0700 (PDT), DKleinecke
> <dkleine...@gmail.com> wrote in
Sounded to me like a suggestion for a differentiation between a
phrasal construction and a non-phrasal one

DKleinecke

unread,
Apr 8, 2013, 9:23:27 PM4/8/13
to
On Apr 7, 4:45 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <8cecc3e3-a1cf-47ea-87ba-28f48a9d4...@g4g2000yqd.googlegroups.com>,
>
>  DKleinecke <dkleine...@gmail.com> wrote:
> > But compositionality is a blunt tool. Even the "meaning" of a verb is
> > elusive but the "meaning" on a particle is often very hard to pin
> > down. The problem is that English verbs seem best handled by
> > describing the structures
> > they appear in in terms of templates.
>
> It seems like you've rediscovered subcategorization, which linguists
> have known about for quite a few decades.
>
> Nathan
>
> --
> Department of Linguistics
> Swarthmore Collegehttp://sanders.phonologist.org/

I know about subcategorization. I thought that, assuming the
categories are things like noun, verb, adjective, etc., subcategories
were subsets of categories - like ditransitive verbs, verbs of motion
and so on.
Things that are handled in modern grammars by making set membership a
feature. Have I misunderstood? That's the normal use of the sub-
prefix. The book I use as a reference (when Google fails me) -
Huddleston's "English Grammar: an outline" (1988) - doesn't mention
either categorization or subcategorization.

I agree that templates can be forced into a features presentation -
just give each different template a different feature. And, as noted,
a features presentation is isomorphic to a subcategorizing
presentation.

At this point consideration of elegance and efficiency enters the
discussion. I think templates are much easier to use.

Templates are not my invention. I have no idea who invented them. I
learned about them from the command processor that comes with the
Inform computer language (which dates from the early 1990's)

Nathan Sanders

unread,
Apr 9, 2013, 12:12:35 AM4/9/13
to
In article
<c9736bfd-b1de-472a...@l5g2000yqe.googlegroups.com>,
DKleinecke <dklei...@gmail.com> wrote:

> On Apr 7, 4:45�pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > In article
> > <8cecc3e3-a1cf-47ea-87ba-28f48a9d4...@g4g2000yqd.googlegroups.com>,
> >
> > �DKleinecke <dkleine...@gmail.com> wrote:
> > > But compositionality is a blunt tool. Even the "meaning" of a verb is
> > > elusive but the "meaning" on a particle is often very hard to pin
> > > down. The problem is that English verbs seem best handled by
> > > describing the structures
> > > they appear in in terms of templates.
> >
> > It seems like you've rediscovered subcategorization, which linguists
> > have known about for quite a few decades.
>
> I know about subcategorization. I thought that, assuming the
> categories are things like noun, verb, adjective, etc., subcategories
> were subsets of categories -

"subcategorization" != "subcategories"

Subcategorization (a.k.a. subcategorization frame) is an ordered list
of the arguments a word takes, including any specific words:

http://en.wikipedia.org/wiki/Subcategorization

DKleinecke

unread,
Apr 9, 2013, 8:31:27 PM4/9/13
to
On Apr 8, 9:12 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <c9736bfd-b1de-472a-81b3-0c139201d...@l5g2000yqe.googlegroups.com>,
Wikipedia: " the ability/necessity for lexical items (usually verbs)
to require/allow the presence and types of the syntactic arguments
with which they co-occur".

That seems to me to be exactly what I described. Your reply, however,
seem to imply that I somehow went wrong. I think the jargon gets
pretty heavy here. A "syntactic argument" is the same as a slot in a
template. Perhaps this is descending to the de gustibus level.

Incidentally I see a need for templates that do not involve verbs -
like the adjectival template "too A to #" where A and # represent
adjectival phrases and infinitive phrases.

Nathan Sanders

unread,
Apr 9, 2013, 8:51:16 PM4/9/13
to
In article
<e30d6a1d-6da3-4142...@qc10g2000pbb.googlegroups.com>,
DKleinecke <dklei...@gmail.com> wrote:

> On Apr 8, 9:12�pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > In article
> > <c9736bfd-b1de-472a-81b3-0c139201d...@l5g2000yqe.googlegroups.com>,
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > �DKleinecke <dkleine...@gmail.com> wrote:
> > > On Apr 7, 4:45�pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > > > In article
> > > > <8cecc3e3-a1cf-47ea-87ba-28f48a9d4...@g4g2000yqd.googlegroups.com>,
> >
> > > > �DKleinecke <dkleine...@gmail.com> wrote:
> > > > > But compositionality is a blunt tool. Even the "meaning" of a verb is
> > > > > elusive but the "meaning" on a particle is often very hard to pin
> > > > > down. The problem is that English verbs seem best handled by
> > > > > describing the structures
> > > > > they appear in in terms of templates.
> >
> > > > It seems like you've rediscovered subcategorization, which linguists
> > > > have known about for quite a few decades.
> >
> > > I know about subcategorization. �I thought that, assuming the
> > > categories are things like noun, verb, adjective, etc., subcategories
> > > were subsets of categories -
> >
> > "subcategorization" != "subcategories"
> >
> > Subcategorization (a.k.a. subcategorization frame) is an ordered list
> > of the arguments a word takes, including any specific words:
> >
> > http://en.wikipedia.org/wiki/Subcategorization
>
> Wikipedia: " the ability/necessity for lexical items (usually verbs)
> to require/allow the presence and types of the syntactic arguments
> with which they co-occur".
>
> That seems to me to be exactly what I described.

Which is why I said you "rediscovered" subcategorization.

> Incidentally I see a need for templates that do not involve verbs -
> like the adjectival template "too A to #" where A and # represent
> adjectival phrases and infinitive phrases.

Yes, subcategorization exists for other lexical items besides verbs
("afraid of", "reliance on", etc.). Again, this is something
linguists have known for decades.

DKleinecke

unread,
Apr 10, 2013, 9:21:35 PM4/10/13
to
On Apr 9, 5:51 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <e30d6a1d-6da3-4142-af7c-f8186f715...@qc10g2000pbb.googlegroups.com>,
I'm not trying to invent new linguistics. I am trying to develop a
useable set of tools for non-statistical parsing. One of the problems
I need to solve is how do I represent the lexicon. Each (small) text
I parse draws on the lexicon and one of the questions I want to be
able to answer will be (when I pool the sub-lexica) will be how often
each template is used. This is still a work in progress. When and if I
am ever satisfied I will announce the fact on sci.lang. Meanwhile I
cannot resist commenting in especially curious things I discover.

For the record: using English is a make-shift. My real target is a
very famous closed corpus - the Quran.

Nathan Sanders

unread,
Apr 10, 2013, 10:26:55 PM4/10/13
to
In article
<1cb03be7-3288-427b...@ut5g2000pbc.googlegroups.com>,
> I'm not trying to invent new linguistics.

I don't mean to suggest that's your goal, but it does appear to be a
possible outcome.

I think you might save yourself some unnecessary wheel-reinvention if
you read up on previous work.

DKleinecke

unread,
Apr 11, 2013, 9:10:00 PM4/11/13
to
On Apr 10, 7:26 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <1cb03be7-3288-427b-9ada-fd7711d51...@ut5g2000pbc.googlegroups.com>,
I find it Interesting that you think there is previous work on this
particular subject. There is the old school parsing method - which
fails almost immediately if you apply it to real text. There is
corpus linguistics which is potentially very interesting but lumps the
language of many speakers together. There have been various attempts
to identify authors and assign authorship using features of particular
author's style.
I have examined all these and none allow me to meet my goal.

The problem is, if I may step outside linguistics for a moment, that
"only scripture interprets scripture". In the case of the Bible, with
a multitude of authors, this can only be done at the semantic level.
The speech of, say, Judges cannot be used to explain the speech of
Nehemiah. But Paul's authentic letters have been separated from non-
authentic by speech (as well as other) criteria. I myself am
reluctant to accept speech patterns as decisive proofs of
authorship.

But the Quran is a fixed closed corpus. There is simply no additional
literature from the same time and place. And, if one were a pious
Muslim, one would say, "Of course, there is nothing comparable to The
Word of Allah". John of Damascus may, or may not, have known of an
additional chapter of the Quran - called The Camel - but in any case
it has vanished without a trace (other than what John says).

For the record I want to state that, while searching for a suitable
already existing framework, the most helpful thing I found was the
methods used by tools for building interactive fictions (text games
such as Zork).

Nathan Sanders

unread,
Apr 11, 2013, 9:39:35 PM4/11/13
to
In article
<1099afa8-4a88-4f20...@gh3g2000pbd.googlegroups.com>,
> I find it Interesting that you think there is previous work on this
> particular subject.

On representing the lexicon? Of course there is.

DKleinecke

unread,
Apr 12, 2013, 7:33:09 PM4/12/13
to
On Apr 11, 6:39 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <1099afa8-4a88-4f20-88a2-9b71e646a...@gh3g2000pbd.googlegroups.com>,
Reference please. The more encyclopedic the better.

Nathan Sanders

unread,
Apr 13, 2013, 2:02:32 PM4/13/13
to
In article
<017ca976-0c25-4623...@kv16g2000pbb.googlegroups.com>,
DKleinecke <dklei...@gmail.com> wrote:

> On Apr 11, 6:39�pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > In article
> > <1099afa8-4a88-4f20-88a2-9b71e646a...@gh3g2000pbd.googlegroups.com>,
> >
> > �DKleinecke <dkleine...@gmail.com> wrote:
> > > On Apr 10, 7:26�pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > > > In article
> > > > <1cb03be7-3288-427b-9ada-fd7711d51...@ut5g2000pbc.googlegroups.com>,
> >
> > > > �DKleinecke <dkleine...@gmail.com> wrote:
> > > > > On Apr 9, 5:51�pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > > > > > In article
> > > > > > <e30d6a1d-6da3-4142-af7c-f8186f715...@qc10g2000pbb.googlegroups.com>
> > > > > > ,
> >
> > > > > > �DKleinecke <dkleine...@gmail.com> wrote:
> > > > > > > On Apr 8, 9:12�pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > > > > > > > In article
> > > > > > > > <c9736bfd-b1de-472a-81b3-0c139201d...@l5g2000yqe.googlegroups.co
> Reference please. The more encyclopedic the better.

http://en.wikipedia.org/wiki/Subcategorization

DKleinecke

unread,
Apr 13, 2013, 9:15:17 PM4/13/13
to
On Apr 13, 11:02 am, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <017ca976-0c25-4623-9b0a-003d1a2b9...@kv16g2000pbb.googlegroups.com>,
First thing I read long ago and what I paraphrased for you a few posts
ago.

The methodology you seem to want to advance must be the one
illustrated there by the examples
work [NP __ ]
eat [NP __ NP]
wait [NP __ for]
These would be written as templates
S work
S eat O
S wait O for
where the third example is missing an additional O.

Which of these two isomorphic ways of presenting the information is
better is a matter of taste - I prefer templates. And, of course, one
cannot use NP for both attributives (or whatever you call them) of
"eat" unless you ignore pronouns as a realization of NP.

I started, several years ago, with this kind of naive representation
and almost immediately found it too rigid to be comfortable to use.

I am not comfortable with the categorization and subcategorization
technical terms because they are Chomskian concepts and I may not be
aware of the nuances they carry.

Another thing worth mentioning is that the template view allows me to
treat idioms and verbal "subcategorizations" as the same phenomenon.
I know of no other systematic way to handle idioms.







Nathan Sanders

unread,
Apr 13, 2013, 11:53:20 PM4/13/13
to
In article
<91378fb3-364f-4d5d...@fz1g2000pbb.googlegroups.com>,
DKleinecke <dklei...@gmail.com> wrote:

> On Apr 13, 11:02�am, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > In article
> > <017ca976-0c25-4623-9b0a-003d1a2b9...@kv16g2000pbb.googlegroups.com>,
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > �DKleinecke <dkleine...@gmail.com> wrote:
> > > On Apr 11, 6:39�pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > > > In article
> > > > <1099afa8-4a88-4f20-88a2-9b71e646a...@gh3g2000pbd.googlegroups.com>,
> >
> > > > �DKleinecke <dkleine...@gmail.com> wrote:
> > > > > On Apr 10, 7:26�pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > > > > > In article
> > > > > > <1cb03be7-3288-427b-9ada-fd7711d51...@ut5g2000pbc.googlegroups.com>,
> >
> > > > > > �DKleinecke <dkleine...@gmail.com> wrote:
> > > > > > > On Apr 9, 5:51�pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > > > > > > > In article
> > > > > > > > <e30d6a1d-6da3-4142-af7c-f8186f715...@qc10g2000pbb.googlegroups.
> > > > > > > > com>
> > > > > > > > ,
> >
> > > > > > > > �DKleinecke <dkleine...@gmail.com> wrote:
> > > > > > > > > On Apr 8, 9:12�pm, Nathan Sanders <sand...@alum.mit.edu>
> > > > > > > > > wrote:
> > > > > > > > > > In article
> > > > > > > > > > <c9736bfd-b1de-472a-81b3-0c139201d...@l5g2000yqe.googlegroup
> Which of these two isomorphic ways of presenting the information is
> better is a matter of taste - I prefer templates.

Of course you do: you are familiar with them; you came up with them.

> And, of course, one
> cannot use NP for both attributives (or whatever you call them) of
> "eat" unless you ignore pronouns as a realization of NP.

The modern label is "DP", which covers not only older-style NPs and
pronouns, but also names. But the specific symbols used for the
notation are irrelevant, as long as the notation allows those three
entities to fill the same syntactic slot.

> Another thing worth mentioning is that the template view allows me to
> treat idioms and verbal "subcategorizations" as the same phenomenon.
> I know of no other systematic way to handle idioms.

Then the two things are not isomorphic, as you claimed above. Either
they can both be used for the exact same set of cases, or they cannot.

DKleinecke

unread,
Apr 14, 2013, 9:19:32 PM4/14/13
to
On Apr 13, 8:53 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <91378fb3-364f-4d5d-81c3-25f1401cf...@fz1g2000pbb.googlegroups.com>,
You are I fear, spending too much time looking for nits. I am writing
casually - not in a pedantic way. In the first paragraph I used
"isomorphic" to describe two ways of representing the lexicon - one of
which was by templates. By the last paragraph I had expanded the use
of templates. The pedantic statement would be that the other form was
isomorphic to a subset of the template forms.

You will have to take up the NP versus DP notation with Wikipedia - I
was quoting Wikipedia. I think the DP theory is funny. It is an
artifact of a theory that demands that each step in a "parse" have a
head and, the noun in the phrase having proved unworkable - I have
forgotten the specific difficulty - they decide that the determiner
must be the head. This is completely counter-intuitive and, in my
opinion, comic. I expect posterity to use it as a parade example of
the nonsense Chomskian thinking drifted off into.

I observe that Fries ("Tagmeme Sequences in the English Noun Phrase"
SIL 36, 1970) has three successive determiner slots and a limiter slot
that precedes the first determiner. None of his examples show all
four slots filled. This seems very diffuse for a head. But then we
already knew there had to be DP with an empty head.

Nathan Sanders

unread,
Apr 14, 2013, 10:17:56 PM4/14/13
to
In article
<d5f0f802-0ab5-4c6f...@aw7g2000pbd.googlegroups.com>,
> You are I fear, spending too much time looking for nits.

You expressed interest in representing the lexicon. I pointed out
that your templates seem to just be notational variants of
subcategorization, which you hadn't heard of. The only "nits" I'm
picking are in helping you to understand what subcategorization is.

You are free to re-invent the wheel if you wish. The major potential
harms in doing so are (1) it will delay your work, and (2) might
introduce errors that could have been avoided had you built upon
pre-existing knowledge instead of starting from scratch. But it's
your baby, so I'm not particularly bothered if your results are
delayed or erroneous.

If your distaste for all things that even vaguely smell of Chomsky is
so overpowering that you're willing to accept delays and errors in
your own work in lexical representation by completely ignoring 50
years worth of work in lexical representations, then that's your
choice to make.

> You will have to take up the NP versus DP notation with Wikipedia - I
> was quoting Wikipedia. I think the DP theory is funny. It is an
> artifact of a theory that demands that each step in a "parse" have a
> head and, the noun in the phrase having proved unworkable -

DP is the result of wanting to treat true noun phrases and pronouns
and names as the same kind of object, within the confines of a
specific theory. My only reason for bringing up DPs was to answer
your objection about pronouns, nothing more. The gross end result of
the DP hypothesis is basically the same as your S/O notation, except
that DP is actually more general, covering both subjects and objects,
which you seem to have no single group notation for.

DKleinecke

unread,
Apr 15, 2013, 8:26:43 PM4/15/13
to
On Apr 14, 7:17 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <d5f0f802-0ab5-4c6f-be63-1f17b952e...@aw7g2000pbd.googlegroups.com>,
Temper, temper, temper. I knew all about subcategorization (and
templates are more than just a notational alternative) but considered
it so obviously inadequate that there was no point in mentioning it.
I see it as one more example of naming something and then offering
that name as an explanation for it..

Why is it desirable to lump subjects and objects when they are
different. Of course the great mass of subjects are also objects and
vice versa. But the totalities are different. Incidentally the pronoun
"you" can be considered part of the greater mass and pronouns "she"
and "it" also behave like the majority but their common plural "they /
them" behaves like a pronoun. There seem to be serious problem in
treating pronouns as a natural class. The best (and it is none too
good) procedure seems to be to set up case sensitive words as a
special class - rather than introducing pronouns as a syntactic
idea..

But back to the start. You have asserted that I am overlooking 50
years of work on representing the lexicon but offered me no references
to any of that work (apart, of course, from the primitive method the
Wikipedia describes). I have told you where I learned about the
template scheme about 25 years ago (thereby falsifying the claim that
I am overlooking everything) - is there, perhaps, no actual work that
I overlooked ?

Nathan Sanders

unread,
Apr 15, 2013, 9:37:05 PM4/15/13
to
In article
<7893a936-3b56-4a93...@id10g2000pbc.googlegroups.com>,
DKleinecke <dklei...@gmail.com> wrote:

> On Apr 14, 7:17锟絧m, Nathan Sanders <sand...@alum.mit.edu> wrote:
>
> > You expressed interest in representing the lexicon. 锟絀 pointed out
> > that your templates seem to just be notational variants of
> > subcategorization, which you hadn't heard of. 锟絋he only "nits" I'm
> > picking are in helping you to understand what subcategorization is.
> >
> > You are free to re-invent the wheel if you wish. 锟絋he major potential
> > harms in doing so are (1) it will delay your work, and (2) might
> > introduce errors that could have been avoided had you built upon
> > pre-existing knowledge instead of starting from scratch. 锟紹ut it's
> > your baby, so I'm not particularly bothered if your results are
> > delayed or erroneous.
> >
> > If your distaste for all things that even vaguely smell of Chomsky is
> > so overpowering that you're willing to accept delays and errors in
> > your own work in lexical representation by completely ignoring 50
> > years worth of work in lexical representations, then that's your
> > choice to make.
> >
> > > You will have to take up the NP versus DP notation with Wikipedia - I
> > > was quoting Wikipedia. I think the DP theory is funny. It is an
> > > artifact of a theory that demands that each step in a "parse" have a
> > > head and, the noun in the phrase having proved unworkable -
> >
> > DP is the result of wanting to treat true noun phrases and pronouns
> > and names as the same kind of object, within the confines of a
> > specific theory. 锟組y only reason for bringing up DPs was to answer
> > your objection about pronouns, nothing more. 锟絋he gross end result of
> > the DP hypothesis is basically the same as your S/O notation, except
> > that DP is actually more general, covering both subjects and objects,
> > which you seem to have no single group notation for.
>
> I knew all about subcategorization

"All about"? Your initial response to my mention of subcategorization
suggests otherwise, as does your follow-up request for an encyclopedic
reference.

If you knew "all about" subcategorization, you would have responded
more appropriately the first time and wouldn't have needed to look at
Wikipedia to understand the meaning of the term "subcategorization",
and you certainly wouldn't have needed to copy their examples, since
anyone who truly knows "all about" subcategorization can trivially
write a subcategorization frame on the fly for nearly any basic verb
in their native language.

> Why is it desirable to lump subjects and objects

Because they have the same internal syntactic structure. Any kind of
phrase that can be a subject can also be an object, and vice versa.

(There are certain possible exceptions, depending on what data you
allow. Overtly case-marked items and reflexive pronouns generally
have distributional restrictions on subject versus object position,
but in my opinion, these seem best handled by the morphology and
semantics, respectively, not by the syntax, despite what Chomsky might
say.)

> when they are different.

They differ in position, not in content or internal structure. That
warrants giving them different position in the
template/subcategorization, but does not warrant assigning them to
inherently different syntactic categories because that implies they
have different syntactic structure.

> But back to the start. You have asserted that I am overlooking 50
> years of work on representing the lexicon

Because you demonstrated that you didn't know what "subcategorization"
is.

> but offered me no references
> to any of that work

Wikipedia has pointers to some original scholarly sources, and I sent
you there. Did I also need to duplicate their words by copying and
pasting their references here? Does your web browser not allow you to
see the list of references cited in Wikipedia articles?

DKleinecke

unread,
Apr 17, 2013, 8:14:28 PM4/17/13
to
On Apr 15, 6:37 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <7893a936-3b56-4a93-9476-1e4a5ff3d...@id10g2000pbc.googlegroups.com>,
>
>
>
>
>
>
>
>
>
>  DKleinecke <dkleine...@gmail.com> wrote:
> > On Apr 14, 7:17 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
>
> > > You expressed interest in representing the lexicon.  I pointed out
> > > that your templates seem to just be notational variants of
> > > subcategorization, which you hadn't heard of.  The only "nits" I'm
> > > picking are in helping you to understand what subcategorization is.
>
> > > You are free to re-invent the wheel if you wish.  The major potential
> > > harms in doing so are (1) it will delay your work, and (2) might
> > > introduce errors that could have been avoided had you built upon
> > > pre-existing knowledge instead of starting from scratch.  But it's
> > > your baby, so I'm not particularly bothered if your results are
> > > delayed or erroneous.
>
> > > If your distaste for all things that even vaguely smell of Chomsky is
> > > so overpowering that you're willing to accept delays and errors in
> > > your own work in lexical representation by completely ignoring 50
> > > years worth of work in lexical representations, then that's your
> > > choice to make.
>
> > > > You will have to take up the NP versus DP notation with Wikipedia - I
> > > > was quoting Wikipedia. I think the DP theory is funny. It is an
> > > > artifact of a theory that demands that each step in a "parse" have a
> > > > head and, the noun in the phrase having proved unworkable -
>
> > > DP is the result of wanting to treat true noun phrases and pronouns
> > > and names as the same kind of object, within the confines of a
> > > specific theory.  My only reason for bringing up DPs was to answer
> > > your objection about pronouns, nothing more.  The gross end result of
I detect a dry well.

In conclusion I note that a hardcore tagmemic treatment (as in the
"Bolivian Indian Grammars" SIL 16) would, omitting some other things,
represent an English kernel sentence as something like
Subj: NP / pro 1/2 + Pred: VP + Obj: NP / pro 1/3
and represent pro 2 as a list of {I, we, he, they, who, ...} and pro 3
as a parallel list {me, us, him, them, whom, ...} leaving pro 1 as
{you, she, it, ... }. This is close to my thinking and a technique I
tried using but found clumsy in practice. The names on the slots are
useful if you are discussing the construction but, in general, not as
helpful as one might imagine.

Brian M. Scott

unread,
Apr 17, 2013, 9:09:44 PM4/17/13
to
On Wed, 17 Apr 2013 17:14:28 -0700 (PDT), DKleinecke
<dklei...@gmail.com> wrote in
<news:fd5335a4-0ce1-4522...@i20g2000pbq.googlegroups.com>
in sci.lang,alt.english.usage,alt.usage.english:

> On Apr 15, 6:37�pm, Nathan Sanders <sand...@alum.mit.edu> wrote:

>> In article
>> <7893a936-3b56-4a93-9476-1e4a5ff3d...@id10g2000pbc.googlegroups.com>,

[...]

>>> but offered me no references to any of that work

>> Wikipedia has pointers to some original scholarly
>> sources, and I sent you there. �Did I also need to
>> duplicate their words by copying and pasting their
>> references here? �Does your web browser not allow you to
>> see the list of references cited in Wikipedia articles?

> I detect a dry well.

You've actually examined the WP references and found them
wanting? Then you should say so. As it is, I strongly
suspect that you've not done so, and that you're just making
a bigger fool of yourself.

[...]

DKleinecke

unread,
Apr 18, 2013, 10:52:25 PM4/18/13
to
On Apr 17, 6:09 pm, "Brian M. Scott" <b.sc...@csuohio.edu> wrote:
> On Wed, 17 Apr 2013 17:14:28 -0700 (PDT), DKleinecke
> <dkleine...@gmail.com> wrote in
I looked over the references when this discussion started. I
recognized most of them. I admit that Burton-Robers was new to me and
I neglected to follow up on him at that time. But Note 5 in the
Wikipedia article indicates that, even though I still have no access
to the actual reference, he has nothing useful to offer me. I cannot
work at all in the context I am trying to work unless the subject is
just as much an attributive of the verb as any object.

Anyway I see nothing in any of these references - even Tesniere - that
helps me. I am trying to work out a PRACTICAL methodology and
"subcategorization" interests me only as a name some people use to
designate a set of facts for which I believe templates are a superior
tool.

Nathan Sanders

unread,
Apr 19, 2013, 12:00:24 AM4/19/13
to
In article
<d38e6fd3-4787-43f2...@tz3g2000pbb.googlegroups.com>,
DKleinecke <dklei...@gmail.com> wrote:

> I cannot
> work at all in the context I am trying to work unless the subject is
> just as much an attributive of the verb as any object.

So why do you give the subject and object completely different labels,
instead of the same label?

Does your "context" require you to redundantly specify that a subject
is both in a subject position and has a label "S" distinct from the
label for objects?

Peter T. Daniels

unread,
Apr 19, 2013, 7:35:17 AM4/19/13
to
On Apr 19, 12:00 am, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <d38e6fd3-4787-43f2-8472-276885c3e...@tz3g2000pbb.googlegroups.com>,
>
>  DKleinecke <dkleine...@gmail.com> wrote:
> > I cannot
> > work at all in the context I am trying to work unless the subject is
> > just as much an attributive of the verb as any object.
>
> So why do you give the subject and object completely different labels,
> instead of the same label?
>
> Does your "context" require you to redundantly specify that a subject
> is both in a subject position and has a label "S" distinct from the
> label for objects?

What do you mean by "subject position"? That's reminiscent of
McCawley's "English as a VSO Language," which seemed to be nothing but
an attempt to express syntactic structure in terms of Reverse Polish
notation. A subject need not have a specific position, even within one
language.

Nathan Sanders

unread,
Apr 19, 2013, 12:35:42 PM4/19/13
to
In article
<ea19e4b5-bda2-4418...@y14g2000vbk.googlegroups.com>,
"Peter T. Daniels" <gram...@verizon.net> wrote:

> On Apr 19, 12:00 am, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > In article
> > <d38e6fd3-4787-43f2-8472-276885c3e...@tz3g2000pbb.googlegroups.com>,
> >
> >  DKleinecke <dkleine...@gmail.com> wrote:
> > > I cannot
> > > work at all in the context I am trying to work unless the subject is
> > > just as much an attributive of the verb as any object.
> >
> > So why do you give the subject and object completely different labels,
> > instead of the same label?
> >
> > Does your "context" require you to redundantly specify that a subject
> > is both in a subject position and has a label "S" distinct from the
> > label for objects?
>
> What do you mean by "subject position"?

The position where "S" occurs in David's templates.

Peter T. Daniels

unread,
Apr 19, 2013, 3:02:08 PM4/19/13
to
On Apr 19, 12:35 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <ea19e4b5-bda2-4418-88b0-893a5757c...@y14g2000vbk.googlegroups.com>,
>  "Peter T. Daniels" <gramma...@verizon.net> wrote:
> > On Apr 19, 12:00 am, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > > In article
> > > <d38e6fd3-4787-43f2-8472-276885c3e...@tz3g2000pbb.googlegroups.com>,
> > >  DKleinecke <dkleine...@gmail.com> wrote:

> > > > I cannot
> > > > work at all in the context I am trying to work unless the subject is
> > > > just as much an attributive of the verb as any object.
>
> > > So why do you give the subject and object completely different labels,
> > > instead of the same label?
>
> > > Does your "context" require you to redundantly specify that a subject
> > > is both in a subject position and has a label "S" distinct from the
> > > label for objects?
>
> > What do you mean by "subject position"?
>
> The position where "S" occurs in David's templates.

I rather doubt that he intends his templates to indicate a specific
order of constituents. It was you, not he, who mentioned "subject
_position_." He used a tagmeme-like notation, displaying a slot-filler
correlation.

Nathan Sanders

unread,
Apr 19, 2013, 3:20:47 PM4/19/13
to
In article
<4e5e83c9-854a-4223...@c9g2000vbr.googlegroups.com>,
"Peter T. Daniels" <gram...@verizon.net> wrote:

> On Apr 19, 12:35 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > In article
> > <ea19e4b5-bda2-4418-88b0-893a5757c...@y14g2000vbk.googlegroups.com>,
> >  "Peter T. Daniels" <gramma...@verizon.net> wrote:
> > > On Apr 19, 12:00 am, Nathan Sanders <sand...@alum.mit.edu> wrote:
> > > > In article
> > > > <d38e6fd3-4787-43f2-8472-276885c3e...@tz3g2000pbb.googlegroups.com>,
> > > >  DKleinecke <dkleine...@gmail.com> wrote:
>
> > > > > I cannot
> > > > > work at all in the context I am trying to work unless the subject is
> > > > > just as much an attributive of the verb as any object.
> >
> > > > So why do you give the subject and object completely different labels,
> > > > instead of the same label?
> >
> > > > Does your "context" require you to redundantly specify that a subject
> > > > is both in a subject position and has a label "S" distinct from the
> > > > label for objects?
> >
> > > What do you mean by "subject position"?
> >
> > The position where "S" occurs in David's templates.
>
> I rather doubt that he intends his templates to indicate a specific
> order of constituents.

So you think the elements of his templates can occur in any order?
That the template "S eat O" could actually mean any of the six
possible orderings are valid?

> It was you, not he, who mentioned "subject
> _position_." He used a tagmeme-like notation, displaying a slot-filler
> correlation.

What is the different between a "slot" and a "position"?

Peter T. Daniels

unread,
Apr 19, 2013, 4:55:54 PM4/19/13
to
On Apr 19, 3:20 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <4e5e83c9-854a-4223-a71a-132380106...@c9g2000vbr.googlegroups.com>,
You never learned anything about tagmemics? "Slot" is not a position,
it's a function.

Nathan Sanders

unread,
Apr 19, 2013, 5:35:20 PM4/19/13
to
In article
<bb5976d1-cffe-4769...@n5g2000yqg.googlegroups.com>,
"Peter T. Daniels" <gram...@verizon.net> wrote:

So the template "S eat O" tells us nothing about word order?

Peter T. Daniels

unread,
Apr 19, 2013, 11:05:23 PM4/19/13
to
On Apr 19, 5:35 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <bb5976d1-cffe-4769-b518-92da2f0ea...@n5g2000yqg.googlegroups.com>,
You'll have to ask David. It certainly doesn't _necessarily_ tell us
anything about word order, since the elements have to be put down in
_some_ order, and the English order is the one that comes most
naturally to the English-speaker.

Nathan Sanders

unread,
Apr 20, 2013, 1:23:47 AM4/20/13
to
In article
<3ded4fe4-5c94-4ba5...@cm2g2000vbb.googlegroups.com>,
"Peter T. Daniels" <gram...@verizon.net> wrote:

So you no longer "doubt that he intends his templates to indicate a
specific order of constituents"?

Peter T. Daniels

unread,
Apr 20, 2013, 7:29:13 AM4/20/13
to
On Apr 20, 1:23 am, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <3ded4fe4-5c94-4ba5-a931-8c620ddd4...@cm2g2000vbb.googlegroups.com>,
How could you possibly get that from what I wrote? (Which, of course,
because you are unable to respond to it, you simply deleted.)

Given everything he has said, there is every reason to suppose that
his main interest is _not_ English, specifically, and therefore what
happens to be the unmarked order for him as an English-speaker says
nothing about his theory as a whole.

johnk

unread,
Apr 20, 2013, 8:18:25 AM4/20/13
to
"Subject position" does not necessarily refer to a linear position. I understood it to mean the position of a subject within a syntactic framework. For instance, old transformational syntax said the subject was the NP directly dominated by the S node while a direct object is the NP directly dominated by the VP. It doesn't matter what the word order is. What Nathan seems to be pointing out is that the NP of a subject is the same as an NP as an object and he is wanting to have the OP explain how he differentiates between S and O in his 'syntactic theory'.
JohnK

Peter T. Daniels

unread,
Apr 20, 2013, 12:46:50 PM4/20/13
to
David, obviously, is not interested in "direct domination"; I expect
his account of Subject and Object (etc.) resembles that of tagmemics.

If that was what Nathan intended, why can't he say so?
--
If you can discover why your newsreader has begun adding lines between
lines you quote, you might be able to help yangg stop doing so.

Nathan Sanders

unread,
Apr 20, 2013, 4:00:28 PM4/20/13
to
In article
<e9b0a5a1-3370-470a...@o9g2000vbk.googlegroups.com>,
"Peter T. Daniels" <gram...@verizon.net> wrote:

Because you told me to ask David, instead of reaffirming your original
doubt about his templates indicating order.

Nathan Sanders

unread,
Apr 20, 2013, 4:00:55 PM4/20/13
to
In article <3ab3a090-a532-45b2...@googlegroups.com>,
Exactly.

Peter T. Daniels

unread,
Apr 20, 2013, 5:03:02 PM4/20/13
to
On Apr 20, 4:00 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article <3ab3a090-a532-45b2...@googlegroups.com>,
Then why didn't you say what you meant? That you were referring to
Chomskyan hocus-pocus rather than "positions" in the sentence, let
alone the actual functions of the sentence components?

Brian M. Scott

unread,
Apr 20, 2013, 5:05:31 PM4/20/13
to
On Sat, 20 Apr 2013 05:18:25 -0700 (PDT), johnk
<jhoba...@gmail.com> wrote in
<news:3ab3a090-a532-45b2...@googlegroups.com>
in sci.lang:

[...]

> What Nathan seems to be pointing out is that the NP of a
> subject is the same as an NP as an object and he is
> wanting to have the OP explain how he differentiates
> between S and O in his 'syntactic theory'.

That OP looks right at home ...

Brian

Nathan Sanders

unread,
Apr 20, 2013, 5:18:29 PM4/20/13
to
In article
<55902f86-5d05-434a...@x14g2000vba.googlegroups.com>,
"Peter T. Daniels" <gram...@verizon.net> wrote:

I have said what I meant.

> That you were referring to Chomskyan hocus-pocus

Huh?

DKleinecke

unread,
Apr 20, 2013, 8:26:52 PM4/20/13
to
On Apr 20, 2:18 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <55902f86-5d05-434a-bd7b-6c7fcd44c...@x14g2000vba.googlegroups.com>,
I think I should speak up. And, of course, Peter is right. I am only
using English as a convenient test case - my real interests lie
elsewhere (Arabic and Native (South) American) languages.

I tried using the same symbol for the fillers of the subject and
object slot (I used N) but it had no advantages and I lost a tiny bit
of information. Real life English has a surprising number of sentences
that are not SVO. When I process beyond the parse I lump S and O
together for lexical purposes.

From my perspective the S slot is the marked one - according to what I
have seen. The default form of an English pronoun is the objective
form - the subject form occurs is one particular usage. So I would say
the English pronouns are "me", "you", "him" ... This is, of course,
the same as the situation in Arabic (where there is also a third
case). But not in many other languages (especially ergatve ones).

In proper tagmemics, especially Fries' less extreme "Tagmemic
Sequences in the English Noun Phrase" (SIL #36), a construction is
displayed as a list of slot names. Then the construction is described
- in plain English - and its order, optionality and repeatability are
described in English. Then the slot fillers are listed slot-by-slot
and exemplified. Some of the slot fillers are themselves constructions
that must be expanded. Fries has fairly lengthy discussion of some of
these situations.

I do not claim to be using tagmemics - only that I am strongly
influenced by tagmemics. I do not, for example, believe that tagmemics
provides an adequate treatment of phrasal verb constructions.

Nathan Sanders

unread,
Apr 20, 2013, 8:33:53 PM4/20/13
to
In article
<47c57958-5a6f-41fe...@ka6g2000pbb.googlegroups.com>,
DKleinecke <dklei...@gmail.com> wrote:

> I tried using the same symbol for the fillers of the subject and
> object slot (I used N) but it had no advantages and I lost a tiny bit
> of information.

Then I still don't understand your original statement that led to this
sub-discussion: "I cannot work at all in the context I am trying to
work unless the subject is just as much an attributive of the verb as
any object."

If you need them to be the same, then why give them different labels?

DKleinecke

unread,
Apr 21, 2013, 8:32:37 PM4/21/13
to
On Apr 20, 5:33 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
> In article
> <47c57958-5a6f-41fe-b97f-a31a9527b...@ka6g2000pbb.googlegroups.com>,
>
>  DKleinecke <dkleine...@gmail.com> wrote:
> > I tried using the same symbol for the fillers of the subject and
> > object slot (I used N) but it had no advantages and I lost a tiny bit
> > of information.
>
> Then I still don't understand your original statement that led to this
> sub-discussion: "I cannot work at all in the context I am trying to
> work unless the subject is just as much an attributive of the verb as
> any object."
>
> If you need them to be the same, then why give them different labels?
>
> Nathan
>
> --
> Department of Linguistics
> Swarthmore Collegehttp://sanders.phonologist.org/

Ah. I will try to answer your question as I think it is meant.

I want the subject and the objects all to be attributives - but
different attributives. There are more kinds of attributives because
each of the traditional adjuncts is also an attributive. For temporal
adjunct I use T and for locatives L and so on. A full sentence parse
might include any of them. For example "Yesterday I saw John at the
supermarket" would end up "T S saw O L".

Thus I am rejecting a fundamental part of x-bar theory - the
distinction between adjuncts and compliments. The reason I am doing so
is because, since I am parsing text and not looking at my own
"competence", I cannot tell - at parse time - which attributives are
optional and which are obligatory. Those facts would have to rise out
of a large corpus.

I cannot use any kind of "competence" or innate knowledge because I am
not a native speaker of any of my target languages.

Nathan Sanders

unread,
Apr 21, 2013, 8:39:17 PM4/21/13
to

Peter T. Daniels

unread,
Apr 21, 2013, 10:37:22 PM4/21/13
to
On Apr 21, 8:32 pm, DKleinecke <dkleine...@gmail.com> wrote:
> On Apr 20, 5:33 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
>
>
>
>
>
>
>
>
>
> > In article
> > <47c57958-5a6f-41fe-b97f-a31a9527b...@ka6g2000pbb.googlegroups.com>,
>
> >  DKleinecke <dkleine...@gmail.com> wrote:
> > > I tried using the same symbol for the fillers of the subject and
> > > object slot (I used N) but it had no advantages and I lost a tiny bit
> > > of information.
>
> > Then I still don't understand your original statement that led to this
> > sub-discussion: "I cannot work at all in the context I am trying to
> > work unless the subject is just as much an attributive of the verb as
> > any object."
>
> > If you need them to be the same, then why give them different labels?
>
> > Nathan
>
> > --
> > Department of Linguistics
> > Swarthmore Collegehttp://sanders.phonologist.org/
>
> Ah. I will try to answer your question as I think it is meant.
>
> I want the subject and the objects all to be attributives - but
> different attributives. There are more kinds of attributives because
> each of the traditional adjuncts is also an attributive.  For temporal
> adjunct I use T and for locatives  L and so on. A full sentence parse
> might include any of them. For example "Yesterday I saw John at the
> supermarket" would end up "T S saw O L".

It sounds a lot like Fillmore's Case Grammar, which ended up (like
Chomsky's, see Baker, Atoms of Language) having to add Case after Case
to account for newly discovered relations.

Ruud Harmsen

unread,
Apr 22, 2013, 3:38:49 AM4/22/13
to
Mon, 15 Apr 2013 17:26:43 -0700 (PDT): DKleinecke
<dklei...@gmail.com> schreef/wrote:

>On Apr 14, 7:17 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
>> In article
>> <d5f0f802-0ab5-4c6f-be63-1f17b952e...@aw7g2000pbd.googlegroups.com>,
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>  DKleinecke <dkleine...@gmail.com> wrote:
>> > On Apr 13, 8:53 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
>> > > In article
>> > > <91378fb3-364f-4d5d-81c3-25f1401cf...@fz1g2000pbb.googlegroups.com>,
>>
>> > >  DKleinecke <dkleine...@gmail.com> wrote:
>> > > > On Apr 13, 11:02 am, Nathan Sanders <sand...@alum.mit.edu> wrote:
>> > > > > In article
>> > > > > <017ca976-0c25-4623-9b0a-003d1a2b9...@kv16g2000pbb.googlegroups.com>,
>>
>> > > > >  DKleinecke <dkleine...@gmail.com> wrote:
>> > > > > > On Apr 11, 6:39 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
>> > > > > > > In article
>> > > > > > > <1099afa8-4a88-4f20-88a2-9b71e646a...@gh3g2000pbd.googlegroups.com>,
>>
>> > > > > > >  DKleinecke <dkleine...@gmail.com> wrote:
>> > > > > > > > On Apr 10, 7:26 pm, Nathan Sanders <sand...@alum.mit.edu> wrote:
>> > > > > > > > > In article
>> > > > > > > > > <1cb03be7-3288-427b-9ada-fd7711d51...@ut5g2000pbc.googlegroups.c
>> > > > > > > > > om>,
>>
>> > > > > > > > >  DKleinecke <dkleine...@gmail.com> wrote:
>> > > > > > > > > > On Apr 9, 5:51 pm, Nathan Sanders <sand...@alum.mit.edu>
>> > > > > > > > > > wrote:
>> > > > > > > > > > > In article
>> > > > > > > > > > > <e30d6a1d-6da3-4142-af7c-f8186f715...@qc10g2000pbb.googlegro
>> > > > > > > > > > > ups.
>> > > > > > > > > > > com>
>> > > > > > > > > > > ,
>>
>> > > > > > > > > > >  DKleinecke <dkleine...@gmail.com> wrote:
>> > > > > > > > > > > > On Apr 8, 9:12 pm, Nathan Sanders <sand...@alum.mit.edu>
>> > > > > > > > > > > > wrote:
>> > > > > > > > > > > > > In article
>> > > > > > > > > > > > > <c9736bfd-b1de-472a-81b3-0c139201d...@l5g2000yqe.googleg
>> > > > > > > > > > > > > roup
>> > > > > > > > > > > > > s.co
>> > > > > > > > > > > > > m>,
>>
>> > > > > > > > > > > > >  DKleinecke <dkleine...@gmail.com> wrote:
>> > > > > > > > > > > > > > On Apr 7, 4:45 pm, Nathan Sanders
>> > > > > > > > > > > > > > <sand...@alum.mit.edu>
>> > > > > > > > > > > > > > wrote:
>> > > > > > > > > > > > > > > In article
>> > > > > > > > > > > > > > > <8cecc3e3-a1cf-47ea-87ba-28f48a9d4...@g4g2000yqd.goo
>> > > > > > > > > > > > > > > gleg
>> > > > > > > > > > > > > > > roup
>> > > > > > > > > > > > > > > s.co
>> > > > > > > > > > > > > > > m>,
>>
>> > > > > > > > > > > > > > >  DKleinecke <dkleine...@gmail.com> wrote:
>> > > > > > > > > > > > > > > > But compositionality is a blunt tool. Even the
>> > > > > > > > > > > > > > > > "meaning"
>> > > > > > > > > > > > > > > > of a
>> > > > > > > > > > > > > > > > verb is
>> > > > > > > > > > > > > > > > elusive but the "meaning" on a particle is often
>> > > > > > > > > > > > > > > > very
>> > > > > > > > > > > > > > > > hard to
>> > > > > > > > > > > > > > > > pin
>> > > > > > > > > > > > > > > > down. The problem is that English verbs seem best
>> > > > > > > > > > > > > > > > handled
>> > > > > > > > > > > > > > > > by
>> > > > > > > > > > > > > > > > describing the structures
>> > > > > > > > > > > > > > > > they appear in in terms of templates.
>>
>> > > > > > > > > > > > > > > It seems like you've rediscovered
>> > > > > > > > > > > > > > > subcategorization,
>> > > > > > > > > > > > > > > which
>> > > > > > > > > > > > > > > linguists
>> > > > > > > > > > > > > > > have known about for quite a few decades.
>>
>> > > > > > > > > > > > > > I know about subcategorization.  I thought that,
>> > > > > > > > > > > > > > assuming
>> > > > > > > > > > > > > > the
>> > > > > > > > > > > > > > categories are things like noun, verb, adjective,
>> > > > > > > > > > > > > > etc.,
>> > > > > > > > > > > > > > subcategories
>> > > > > > > > > > > > > > were subsets of categories -
>>
>> > > > > > > > > > > > > "subcategorization" != "subcategories"
>>
>> > > > > > > > > > > > > Subcategorization (a.k.a. subcategorization frame) is
>> > > > > > > > > > > > > an
>> > > > > > > > > > > > > ordered
>> > > > > > > > > > > > > list
>> > > > > > > > > > > > > of the arguments a word takes, including any specific
>> > > > > > > > > > > > > words:
>>
>> > > > > > > > > > > > >http://en.wikipedia.org/wiki/Subcategorization
>>
>> > > > > > > > > > > > Wikipedia: " the ability/necessity for lexical items
>> > > > > > > > > > > > (usually
>> > > > > > > > > > > > verbs)
>> > > > > > > > > > > > to require/allow the presence and types of the syntactic
>> > > > > > > > > > > > arguments
>> > > > > > > > > > > > with which they co-occur".
>>
>> > > > > > > > > > > > That seems to me to be exactly what I described.
>>
>> > > > > > > > > > > Which is why I said you "rediscovered" subcategorization.
>>
>> > > > > > > > > > > > Incidentally I see a need for templates that do not
>> > > > > > > > > > > > involve
>> > > > > > > > > > > > verbs
>> > > > > > > > > > > > -
>> > > > > > > > > > > > like the adjectival template "too A to #" where A and #
>> > > > > > > > > > > > represent
>> > > > > > > > > > > > adjectival phrases and infinitive phrases.
>>
>> > > > > > > > > > > Yes, subcategorization exists for other lexical items
>> > > > > > > > > > > besides
>> > > > > > > > > > > verbs
>> > > > > > > > > > > ("afraid of", "reliance on", etc.).  Again, this is
>> > > > > > > > > > > something
>> > > > > > > > > > > linguists have known for decades.
>>
>> > > > > > > > > > I'm not trying to invent new linguistics.
>>
>> > > > > > > > > I don't mean to suggest that's your goal, but it does appear to
>> > > > > > > > > be
>> > > > > > > > > a
>> > > > > > > > > possible outcome.
>>
>> > > > > > > > > I think you might save yourself some unnecessary
>> > > > > > > > > wheel-reinvention
>> > > > > > > > > if
>> > > > > > > > > you read up on previous work.
>>
>> > > > > > > > I find it Interesting that you think there is previous work on
>> > > > > > > > this
>> > > > > > > > particular subject.
>>
>> > > > > > > On representing the lexicon?  Of course there is.
>>
>> > > > > > Reference please. The more encyclopedic the better.
>>
>> > > > >http://en.wikipedia.org/wiki/Subcategorization
>>
>> > > > Which of these two isomorphic ways of presenting the information is
>> > > > better is a matter of taste - I prefer templates.
>>
>> > > Of course you do: you are familiar with them; you came up with them.
>>
>> > > > And, of course, one
>> > > > cannot use NP for both attributives (or  whatever you call them) of
>> > > > "eat" unless you ignore pronouns as a realization of NP.
>>
>> > > The modern label is "DP", which covers not only older-style NPs and
>> > > pronouns, but also names.  But the specific symbols used for the
>> > > notation are irrelevant, as long as the notation allows those three
>> > > entities to fill the same syntactic slot.
>>
>> > > > Another thing worth mentioning is that the template view allows me to
>> > > > treat idioms and verbal "subcategorizations" as the same phenomenon.
>> > > > I know of no other systematic way to handle idioms.
>>
>> > > Then the two things are not isomorphic, as you claimed above.  Either
>> > > they can both be used for the exact same set of cases, or they cannot.
>>
>> > You are I fear, spending too much time looking for nits.
>> Nathan
>>
>> --
>> Department of Linguistics
>> Swarthmore Collegehttp://sanders.phonologist.org/


Er mag weer geknipt worden, mensen!!

DKleinecke

unread,
Apr 22, 2013, 10:01:08 PM4/22/13
to
On Apr 21, 7:37 pm, "Peter T. Daniels" <gramma...@verizon.net> wrote:

> It sounds a lot like Fillmore's Case Grammar, which ended up (like
> Chomsky's, see Baker, Atoms of Language) having to add Case after Case
> to account for newly discovered relations.
>

I have decided to speak of what I used to call "slot" under the name
"function" (Fries' nomenclature)

I would not call them "cases" unless I could identify each lexical
entry with a set of differing forms across different functions. We all
know examples. We also know that it is possible that the same case
form will be used in multiple functions - so there would be three
stages (1) lexical entry (2) case (3) function. And, of course, I am
referring to nouns. English case is just barely worth discussing. I
believe that definition makes gender and number aspects of case but I
haven't thought that through.

And yes I anticipate a large number of categories (parts of speech).
To the point of perhaps not naming them. I use a very shortened list
of parts of speech currently in order to get things done and I have an
X category for things that are just too confusing. I am not currently
very careful about anything except noun, verb and adjective
phrases.

I do think Fillmore got a little carried away. But Case Grammar (and
theme's) do provide nice lists of what one must consider.

Peter T. Daniels

unread,
Apr 22, 2013, 11:15:03 PM4/22/13
to
On Apr 22, 10:01 pm, DKleinecke <dkleine...@gmail.com> wrote:
> On Apr 21, 7:37 pm, "Peter T. Daniels" <gramma...@verizon.net> wrote:
>
> > It sounds a lot like Fillmore's Case Grammar, which ended up (like
> > Chomsky's, see Baker, Atoms of Language) having to add Case after Case
> > to account for newly discovered relations.
>
> I have decided to speak of what I used to call "slot" under the name
> "function" (Fries' nomenclature)
>
> I would not call them "cases" unless I could identify each lexical
> entry with a set of differing forms across different functions. We all
> know examples. We also know that it is possible that the same case
> form will be used in multiple functions - so there would be three
> stages (1) lexical entry (2) case (3) function.  And, of course, I am
> referring to nouns. English case is just barely worth discussing. I
> believe that definition makes gender and number aspects of case but I
> haven't thought that through.

Is it coincidence that the three Cambridge Red books on those topics
are all by Greville Corbett (IIRC), who happens to be a Slavicist? And
also the more recent *Agreement*?

> And yes I anticipate a large number of categories (parts of speech).
> To the point of perhaps not naming them. I use a very shortened list
> of parts of speech currently in order to get things done and I have an
> X category for things that are just too confusing.  I am not currently
> very careful about anything except noun, verb and adjective
> phrases.

What's the definition of "part of speech"? Fries et al. used it for
inflectional classes -- Eng. nouns have plural & possessive, verbs
have the 5 forms, adjectives have -er -est.

> I do think Fillmore got a little carried away. But Case Grammar (and
> theme's) do provide nice lists of what one must consider.

Open-ended lists.

DKleinecke

unread,
Apr 23, 2013, 6:51:02 PM4/23/13
to
I don't think there is an agreed-upon definition of "parts of speech".
Huddleston (1988) has eight - Verb, Noun, Adjective, Adverb,
Preposition, Determinative, Coordinator (and, or), Subordinator (that,
whether) - but I am sure he has changed this list several times since
then. I personally don't think the concept is of any use these days
even though it was big before modern linguistics started.

The basic idea is a a crude kind of categorization. I don't know which
Fries et al you mean but I have read none of them so I can't comment
on the usage you quote other than to say it is not the usual meaning
of "part of speech". The vocabulary for the what you quote them as
saying is in no better state. It should be pretty obvious that an
entry in the lexicon is not a single word (except maybe in Chinese)
but a vector of words. Different parts of speech as different vectors
- verbs (except the copula) are five long, adjective (probably should
be adjective 1 as opposed to other adjectives) are three long and
nouns are two. The copula and the pronoun have special vectors all of
their own. This encapsulated the meaning of "inflection" - the idea
being that all these words have the same meaning and differ only
syntactically - but that is a strange kind of meaning that ignores
number, gender and so on. This is just one of many problems with the
lexicon in the treatments - new and old - that I have seen.

People outside the mainstream (for example the authors of the Nuance
Developer's Guide) continue to use the idea of "parts of speech" as
though it explained everything.
Message has been deleted

Simon G

unread,
May 19, 2013, 12:11:51 PM5/19/13
to
Would people agree with my analysis of the following in terms of clause structure (A=adverbial, C=complement)?


They put the presents under the tree 1)

--SVOC. 'They put the presents' is ungrammatical.


She brought in the shopping 2)

SVOA because ‘she brought the shopping’ is grammatical. 'in' used with simple spatial meaning, contrasting with out/up/down etc.


They took in the refugees 3)

SVO with phrasal verb. Idiomatic meaning.





Simon

benl...@ihug.co.nz

unread,
May 19, 2013, 4:43:30 PM5/19/13
to
On May 20, 4:11 am, Simon G <sherringb...@hotmail.com> wrote:
> Would people agree with my analysis of the following in terms of clause structure (A=adverbial, C=complement)?
>
> They put the presents under the tree                1)
>
> --SVOC. 'They put the presents' is ungrammatical.

OK

>
> She brought in the shopping                        2)
>
> SVOA because ‘she brought the shopping’ is grammatical. 'in' used with simple spatial meaning, contrasting with out/up/down etc.

OK, assuming you mean "SVAO".

>
> They took in the refugees                        3)
>
> SVO with phrasal verb. Idiomatic meaning.
>
> Simon

I would say SVAO, i.e. in terms of syntactic categories (2) and (3)
are the same.
The only difference is in the idiomatic meaning associated with "take
in" in (3).

Odysseus

unread,
May 19, 2013, 10:22:47 PM5/19/13
to
In article <4fce9656-d4a8-4583...@googlegroups.com>,
If there's an implied claim that "They took the refugees" is not a
grammatical sentence, I don't see why not -- and given a suitable
context, I think it could be a perfectly natural utterance.

--
Odysseus

Simon G

unread,
May 20, 2013, 12:28:11 PM5/20/13
to
But that implies 'in' and the vb. don't form a unit. Can't we view 'in' as something akin to the separable prefixes of German verbs?

Simon

benl...@ihug.co.nz

unread,
May 20, 2013, 4:31:50 PM5/20/13
to
How does it imply that? It's perfectly normal for groups of words to
have idiomatic meanings, and that's all the unity that "take in" has.

>Can't we view 'in' as something akin to the separable prefixes of German verbs?

We can, in the sense that they are both V + X comprising a lexical/
semantic unit, where X may or may not be separated from the verb. But
what follows from this comparison? I don't see any reason to consider
"in" an affix in the above examples.

Ross Clark
0 new messages