Prepositional phrases as scopal arguments

SENT: A man is spreading butter on a piece of bread

Below I present the top 2 analyses for this sentence.

In the first one, the ARG3 of _spread_v_1 is an h variable. Corresponding to the predication _on_p_loc in a fully resolved scoped tree.

In the second, _spread_v_1 (same sense) has two arguments only. The _on_p_state is now the predicate that connects the event e2 (spread) with the location _piece_n_of.

The readings are challenging to differentiate, but I can understand them. Nevertheless, why the first MRS can’t take, instead of the h9 argument (scopal, right?), the intrinsic variable e15 from the predication _on_p_loc? I suspect that according to a similar discussion here, _on_p_loc and _spread_v_1 would need to share a label, right? Why is that approach not possible here?

The second reading was produced by the rule HD-A_INT-UNSL_C when combining the VP ‘spreading butter’ with the PP ‘on a piece of bread’ . The first one was produced by the rule HD-CMP_U_C taking the same two VP and PP. What ‘foll.int.adjct, no gap’ means?

hd-aj_int-unsl_c := hadj_i_unsl_rule &
"""
Hd+foll.int.adjct, no gap

<ex>B [left quietly].
"""
  [ RNAME hmiu ].

...

hd-cmp_u_c := hcomp_rule &
"""
Hd+complement
<ex>B [hired C].
"""
  [ RNAME hcp ].
SENT: A man is spreading butter on a piece of bread
[ LTOP: h0
INDEX: e2 [ e SF: prop TENSE: pres MOOD: indicative PROG: + PERF: - ]
RELS: < [ _a_q<0:1> LBL: h4 ARG0: x3 [ x PERS: 3 NUM: sg IND: + ] RSTR: h5 BODY: h6 ]
 [ _man_n_1<2:5> LBL: h7 ARG0: x3 ]
 [ _spread_v_1<9:18> LBL: h1 ARG0: e2 ARG1: x3 ARG2: x8 [ x PERS: 3 NUM: sg GEND: n IND: - ] ARG3: h9 ]
 [ udef_q<19:25> LBL: h10 ARG0: x8 RSTR: h11 BODY: h12 ]
 [ _butter_n_1<19:25> LBL: h13 ARG0: x8 ]
 [ _on_p_loc<26:28> LBL: h14 ARG0: e15 [ e SF: prop TENSE: untensed MOOD: indicative PROG: - PERF: - ] ARG1: x8 ARG2: x16 [ x PERS: 3 NUM: sg IND: + ] ]
 [ _a_q<29:30> LBL: h17 ARG0: x16 RSTR: h18 BODY: h19 ]
 [ _piece_n_of-2<31:36> LBL: h20 ARG0: x16 ARG1: x21 [ x PERS: 3 NUM: sg ] ]
 [ udef_q<40:45> LBL: h22 ARG0: x21 RSTR: h23 BODY: h24 ]
 [ _bread_n_1<40:45> LBL: h25 ARG0: x21 ] >
HCONS: < h0 qeq h1 h5 qeq h7 h9 qeq h14 h11 qeq h13 h18 qeq h20 h23 qeq h25 >
ICONS: < > ]

[ LTOP: h0
INDEX: e2 [ e SF: prop TENSE: pres MOOD: indicative PROG: + PERF: - ]
RELS: < [ _a_q<0:1> LBL: h4 ARG0: x3 [ x PERS: 3 NUM: sg IND: + ] RSTR: h5 BODY: h6 ]
 [ _man_n_1<2:5> LBL: h7 ARG0: x3 ]
 [ _spread_v_1<9:18> LBL: h1 ARG0: e2 ARG1: x3 ARG2: x8 [ x PERS: 3 NUM: sg GEND: n IND: - ] ]
 [ udef_q<19:25> LBL: h9 ARG0: x8 RSTR: h10 BODY: h11 ]
 [ _butter_n_1<19:25> LBL: h12 ARG0: x8 ]
 [ _on_p_state<26:28> LBL: h1 ARG0: e13 [ e SF: prop TENSE: untensed MOOD: indicative PROG: - PERF: - ] ARG1: e2 ARG2: x14 [ x PERS: 3 NUM: sg IND: + ] ]
 [ _a_q<29:30> LBL: h15 ARG0: x14 RSTR: h16 BODY: h17 ]
 [ _piece_n_of-2<31:36> LBL: h18 ARG0: x14 ARG1: x19 [ x PERS: 3 NUM: sg ] ]
 [ udef_q<40:45> LBL: h20 ARG0: x19 RSTR: h21 BODY: h22 ]
 [ _bread_n_1<40:45> LBL: h23 ARG0: x19 ] >
HCONS: < h0 qeq h1 h5 qeq h7 h10 qeq h12 h16 qeq h18 h21 qeq h23 >
ICONS: < > ]

I think resultatives are analysed using handle-valued arguments so that scopal modifiers can intervene, e.g. “A man is spreading butter not on a piece of bread but on a banana.”

1 Like

Fantastic! It does make sense. But… Of course, I could not hold myself, and I parsed your example with ERG. 815 readings, but the first is weird for me; it omitted the not.

Does it make sense? @Dan ?

SENT: A man is spreading butter not on a piece of bread but on a banana.

[ LTOP: h0
INDEX: e2 [ e SF: prop TENSE: pres MOOD: indicative PROG: + PERF: - ]
RELS: < [ _a_q<0:1> LBL: h4 ARG0: x3 [ x PERS: 3 NUM: sg IND: + ] RSTR: h5 BODY: h6 ]
 [ _man_n_1<2:5> LBL: h7 ARG0: x3 ]
 [ _spread_v_1<9:18> LBL: h1 ARG0: e2 ARG1: x3 ARG2: x8 [ x PERS: 3 NUM: sg GEND: n IND: - ] ARG3: h9 ]
 [ udef_q<19:25> LBL: h10 ARG0: x8 RSTR: h11 BODY: h12 ]
 [ _butter_n_1<19:25> LBL: h13 ARG0: x8 ]
 [ _on_p_loc<30:32> LBL: h14 ARG0: e15 [ e SF: prop TENSE: untensed MOOD: indicative PROG: - PERF: - ] ARG1: x8 ARG2: x
16 [ x PERS: 3 NUM: sg IND: + ] ]
 [ _a_q<33:34> LBL: h17 ARG0: x16 RSTR: h18 BODY: h19 ]
 [ _piece_n_of-2<35:40> LBL: h20 ARG0: x16 ARG1: x21 [ x PERS: 3 NUM: sg ] ]
 [ udef_q<44:49> LBL: h22 ARG0: x21 RSTR: h23 BODY: h24 ]
 [ _bread_n_1<44:49> LBL: h25 ARG0: x21 ]
 [ _but_c<50:53> LBL: h14 ARG0: e26 [ e SF: prop TENSE: untensed MOOD: indicative PROG: - PERF: - ] ARG1: e15 ARG2: e27
 [ e SF: prop TENSE: untensed MOOD: indicative PROG: - PERF: - ] ]
 [ _on_p_loc<54:56> LBL: h14 ARG0: e27 ARG1: x8 ARG2: x28 [ x PERS: 3 NUM: sg IND: + ] ]
 [ _a_q<57:58> LBL: h29 ARG0: x28 RSTR: h30 BODY: h31 ]
 [ _banana_n_1<59:65> LBL: h32 ARG0: x28 ] >
HCONS: < h0 qeq h1 h5 qeq h7 h9 qeq h14 h11 qeq h13 h18 qeq h20 h23 qeq h25 h30 qeq h32 >
ICONS: < > ]

How strange! Without the coordination (“A man is spreading butter not on a piece of bread”), the ERG gives the analysis I was expecting.

If I try a similar coordination without the resultative (“we saw not cats but dogs”), the ERG uses a special coordination predicate _but_c_not.

So I wonder if the two constructions are not interacting correctly. For the cat-dog sentence and the bread-banana sentence, “not” has type not_mrk, but for the bread-only sentence, it has type not_prdp.

2 Likes