"together" and "separately" don't get a similar reading in the same sentence

grammars ERG
#1

When I parse “which 2 files separately are in this folder?” with ERG2020 I get (among others) the parse below that has the _separate_a_1(ie) predication modifying the introduced variable of _in_p_loc. I (hopefully correctly) interpret this to mean “in separately”:

[ "which 2 files separately are in this folder"
  TOP: h0
  INDEX: e2 [ e SF: ques TENSE: pres MOOD: indicative PROG: - PERF: - ]
  RELS: < [ _which_q<0:5> LBL: h4 ARG0: x3 [ x PERS: 3 NUM: pl IND: + ] RSTR: h5 BODY: h6 ]
          [ card<6:7> LBL: h7 ARG0: e9 [ e SF: prop TENSE: untensed MOOD: indicative PROG: bool PERF: - ] ARG1: x3 CARG: "2" ]
          [ _file_n_of<8:13> LBL: h7 ARG0: x3 ARG1: i10 ]
          [ _separate_a_1<14:24> LBL: h1 ARG0: i11 ARG1: e2 ]
          [ _in_p_loc<29:31> LBL: h1 ARG0: e2 ARG1: x3 ARG2: x12 [ x PERS: 3 NUM: sg IND: + ] ]
          [ _this_q_dem<32:36> LBL: h13 ARG0: x12 RSTR: h14 BODY: h15 ]
          [ _folder_n_of<37:43> LBL: h16 ARG0: x12 ARG1: i17 ] >
  HCONS: < h0 qeq h1 h5 qeq h7 h14 qeq h16 > ]

                                                    ┌── _file_n_of(x3,i10)
                                        ┌────── and(0,1)
                  ┌────── _folder_n_of(x│2,i17)       └ card(2,e9,x3)
_this_q_dem(x12,RSTR,BODY)              │
                       └─ _which_q(x3,RSTR,BODY)
                                             │      ┌── _separate_a_1(i11,e2)
                                             └─ and(0,1)
                                                      └ _in_p_loc(e2,x3,x12)

But with the phrase “which two files together are in this folder”, I get 6 parses, only one of which uses in_p_loc as the “verb” (i.e. instead of be_v), and none of them have a form of “together” which references the index verb’s event variable as _separate_a_1 does above. They all use _together_p(ex). Below is the parse that is closest to the above example. Note that “which files simultaneously/jointly are in this folder?” do produce one in the form I’m expecting.

Two questions:

  • Should there be an analogous parse to the above?
  • Is _together_p(ex) acting like a kind of cardinal in the below phrase? Similiar to the way little-few_a(ex) or _several_a_1(ex) work? Or how should this be interpreted?
[ "which two files together are in this folder"
  TOP: h0
  INDEX: e2 [ e SF: ques TENSE: pres MOOD: indicative PROG: - PERF: - ]
  RELS: < [ _which_q<0:5> LBL: h4 ARG0: x3 [ x PERS: 3 NUM: pl IND: + ] RSTR: h5 BODY: h6 ]
          [ card<6:9> LBL: h7 ARG0: e9 [ e SF: prop TENSE: untensed MOOD: indicative PROG: bool PERF: - ] ARG1: x3 CARG: "2" ]
          [ _file_n_of<10:15> LBL: h7 ARG0: x3 ARG1: i10 ]
          [ _together_p<16:24> LBL: h7 ARG0: e11 [ e SF: prop TENSE: untensed MOOD: indicative PROG: - PERF: - ] ARG1: x3 ]
          [ _in_p_loc<29:31> LBL: h1 ARG0: e2 ARG1: x3 ARG2: x12 [ x PERS: 3 NUM: sg IND: + ] ]
          [ _this_q_dem<32:36> LBL: h13 ARG0: x12 RSTR: h14 BODY: h15 ]
          [ _folder_n_of<37:43> LBL: h16 ARG0: x12 ARG1: i17 ] >
  HCONS: < h0 qeq h1 h5 qeq h7 h14 qeq h16 > ]

-- Parse #1, Tree #0: 

                                                    ┌──── _together_p(e11,x3)
                                                    │ ┌── _file_n_of(x3,i10)
                                        ┌────── and(0,1,2)
                  ┌────── _folder_n_of(x│2,i17)         │
                  │                     │               └ card(2,e9,x3)
_this_q_dem(x12,RSTR,BODY)              │
                       └─ _which_q(x3,RSTR,BODY)
                                             └─ _in_p_loc(e2,x3,x12)
#2

I’m having a hard time interpreting those sentences. Can you give a context that supports the reading you are looking for?

As a side note: the be that goes with PP predicates should be semantically empty, so we wouldn’t expect _be_v in the MRS for either sentence; _in_p_loc seems right.

There are probably two possibilities for how we could analyze together and separately – as modifiers of the subject or as modifiers of the predicate. Since separately is an adverb, it’s pretty clearly going to modify the predicate. together is less immediately clear – but again, I’m not sure what these sentences are meant to mean!

1 Like
#3

I’m implementing collective and distributive readings of cardinals like “two files are in this folder” and I’m now trying to add support for words like “together” and “separate” to force collective or distributive readings.

The collective and distributive readings for the phrases above don’t do anything differently for files in one folder, but I believe they are proper English that someone might say, but unlikely. I’m just using them as degenerate cases to explore how the ERG parses words like “together/separately” when used in various phrases, and what phenomenas appear in the MRS.

A better version of my “degenerate” example that actually has meaningful collectivity is: “two children together are eating a pizza”. And, like my original “degenerate” file example, it only uses the predicate _together_p(e,x), while “two children separately are eating a pizza” uses the _separate_a_1(i,e2), _eat_v_1(e2,x3,x11) construction.

A few more likely phrases about files where the coll/dist readings actually matter are:

“which two files in this folder total 1 megabyte”: (could be either. distributive: 2 files that are each a megabyte, collective: 2 files that add up to a megabyte)
“which two files in this folder total together 1 megabyte”: (forces collective)
“which two files in this folder separately total 1 megabyte”: (forces distributive)

In these phrases, I get the phenomena I am looking for in both the “together” phrase: _together_p_state(e20,e2), _total_v_1(e2,x5,i19) and the “separate” phrase: _separate_a_1(i11,e2), _total_v_1(e2,x3,x19)). But in my simpler pizza and file examples, I only get it for the “separate” phrase.

My hope with all of these examples was that the ERG would always have a reading where “separate/together” were modifying the introduced variable of one of the predications. It seemed like the right interpretation based on what I know and (as a bonus) I understand how to implement it.

So, that’s why I’m wondering if this is just a missing case in the ERG. And, also I’m curious how to interpret that _together_p(e,x) predication that is generated for “two children together are eating a pizza”.

[Edit: I fixed up “the total together” and “separately total” examples. They had a couple of things wrong]
[Edit 2: After more playing around I think I understand that _together_p(e,x) is used when “together” is being interpreted as applying to the noun phrase, but I’m still curious why I’m not getting a together_p_state alternative where it applies to the verb phrase]

#4

Are you in fact getting a syntax tree which has together as a modifier of the VP for “The children together are eating a pizza”? (I’d say that as “The children are eating a pizza together.”)

#5

No. “The children together are eating a pizza” only uses _together_p(e8,x3), _child_n_1(x3) in its two parses.

“The children are eating a pizza together” does get _together_p_state(e15,e2), mod(e9,e2,x10), _eat_v_1(e2,x3,i8). But I’m unclear what mod is trying to say…

[Edit: Correction, there is a parse without the mod predication that is just: _together_p_state(e15,e2), _eat_v_1(e2,x3,i8)]

Agreed that the “x together” construct feels a little cumbersome…