Jon Sung
76-451
Final Project
Get Stuffed! Findings on the “Get +” Construction
But what does that really mean? How can someone “get” stuffed? “Stuffed” is not an object to be retrieved; it’s an adjective! What about “get lost” or “get away”? “Away” is a preposition, which may be even worse. We all know what someone means when s/he says s/he “got up” or “got laid,” but how did that construction come to be?
The question of origin probably can’t be answered by looking in a corpus, but the following questions might:
Has the “got + JJ” or “got + PREP” construction made it into writing? If so, where does it appear? Which corpus is it more common in?
What sorts of adjectives and prepositions are used in this construction? Is there a pattern that can be determined?
Is there some overarching theme or schema used unconsciously in conjunction with this construction? If so, what is it?
If these questions can be answered, they could help to build a better picture of this construction – which, while commonplace, is actually somewhat nonsensical when you really look at it. While it is possible that “get” has become one of those verbs used as a place-holder, like “take” (“take a shower,” “take a sip”, etc.), it is my contention that there exists a structure to its use and collocations.
In searching all corpora in the Cobuild corpus for the following basic constructions:
“get+ VERB”
“get + JJ”
“get+IN”
I discovered that by far, the UKSPOK corpus contains the most instances. The percent difference between occurrences for UKSPOK and the next-highest corpus are illustrated below:
Construction |
Next-Highest Corpus |
Percent Diff |
“get+VERB” |
SUNNOW |
64.98 |
“get+JJ” |
NPR |
72.44 |
“get+IN” |
SUNNOW |
59.73 |
In addition, the SUNNOW corpus appears in the top three for all searches, followed by NPR (with two out of three).
Total number of instances of each of the three forms is as follows:
Construction |
Number of Occurrences |
“get+VERB” |
5960 |
“get+JJ” |
5072 |
“get+IN” |
14314 |
The “get+IN” construction far and away appears to be the most popular, but it presents certain problems which are discussed further on.
Searching the UKSPOK corpus for any kind of pattern with the “get +” construction yielded a peculiar trend: for some reason, adjectives and verbs describing negative actions or situations appear more frequently than those describing positive ones. A sampling of the results from the UKSPOK corpus to the query “get@ + VERB” (there are 24 total):
ukspok/04 op. He makes me listen. He's got learning.
But no tattoos. He's got
ukspok/04
<ZF1> we've got <ZF0> we've got broken slabs in the back
yard as it
ukspok/04 t the seventies when I first got married
you know <M01> Yeah. <M07> I'
ukspok/04 sed to love the way children got chucked
out halfway through the
ukspok/04 go
on the buses and they've got fucking their breasts are down to
ukspok/04 or whatever it is that she's got left to
wait there. But in fact that
ukspok/04 ecided <ZGY> <F0X> FX you've
got `Publishing programme firm dates”.
ukspok/04 st flat together before they got married
and mum wasn't too happy
ukspok/04
field and er as more things got written I was only faced with
ukspok/04
she said to me once 'cos we got chatting about it that she would
ukspok/04 ays I'm carrying and haven't got rid of
and the anger. So then I
ukspok/04 is wife and a friend when he got trapped
in avalanche. Labour says it'
ukspok/04
three days the next. But we got paid. So that last time was the
ukspok/04 d if you weren't careful you got caught in
that. But later on <ZF1>
ukspok/04 rtant.
<M01> That's where I got hit by a car. See that on my knee
ukspok/04
after I'd <ZGY> the I B P I got roped in in nineteen-sixty-three.
ukspok/04
could actually get a j if I got hope f well maybe a an assignment or
ukspok/04 > I see. <M01> And I sort of got
stuck into doing it about mountain
ukspok/04 dmitted that but they a they got dropped
<M01> How was it understood
ukspok/04
weeks got adjourned then it got adjourned for four weeks then after
ukspok/04 nas at him. <M01> No <ZGY> I
got thrown things at me last night.
ukspok/04 en It took me back to when I got married
and <ZGY> <F01>
ukspok/04 to
live? <M02> Mm. After we got married I was living not too far
ukspok/04 ally that they <ZF1> g <ZF0>
got offered better jobs and <F01> Mm.
Counting constructions like “got married” as positive, “got trapped” as negative, and constructions like “got written” as neutral, the sample breaks down as follows:
Connotation |
Number |
Percentage |
Positive |
7 |
29.17 |
Negative |
12 |
50.00 |
Tagging errors/neutral |
5 |
20.83 |
Next, I searched the UKSPOK corpus collocation listings. The node word was “get,” the span 0:1, the t-scores recorded from the top at 36.576064 to 2.200384 at the bottom. There were 72 constructions that described something using an adjective, a verb, or a preposition.[1] The same trend is noted.
Connotation |
Number |
Percentage |
Positive |
19 |
26.39 |
Negative |
41 |
56.94 |
Tagging errors/neutral |
12 |
16.67 |
There is, of course, a slight problem with the neutral constructions, because depending on the context of the sentences they were located in, they could be either good or bad. To search the corpus for each construction and catalogue its occurrences for positive or negative connotations would be an effort entirely too time-consuming, unfortunately, so the margin of error provided by the neutral constructions will have to stand. As it is, it presents a possibility of error that breaks down thusly:
All neutrals shifted to: |
New Percentage |
Margin of Error |
Positive |
43.06 |
24.00 |
Negative |
73.61 |
12.77 |
Additionally, the use of the preposition in the “get +” construction has proven to be a bit less cooperative in an analytical sense, since they turn out almost always to be neutral, and are removed in the sifting process. A context-sensitive look at constructions like “get away” or “get out” may prove to be useful and interesting, but beyond the scope of this analysis, which is more concerned with general trends in the construction.
Searching the collocation t-scores in the same way using the node word “getting” in the UKSPOK corpus recovered some slightly different results. The t-score range here was from 13.779412 to 1.935314. After sifting out the likely constructions, the following numbers were obtained:[2]
Connotation |
Number |
Percentage |
Positive |
8 |
27.58 |
Negative |
12 |
41.38 |
Tagging errors/neutral |
9 |
31.03 |
These numbers are somewhat less useful, since the neutral constructions actually outnumber the positive ones, and the addition of all the neutral tokens to either the positive or negative types could upset the percentages by a significant margin.
A collocation search using the node word “got” in the UKSPOK corpus yielded results more similar to the first search. The t-score range here was from 61.656330 to 2.025026. After removing useless examples, the breakdown looks like this:[3]
Connotation |
Number |
Percentage |
Positive |
11 |
27.50 |
Negative |
23 |
57.50 |
Tagging errors/neutral |
6 |
15.00 |
Once again we see almost twice the number of negative connotations as positive.
If we take into account the possibility that the neutral examples may be either all positive or all negative, the following percentages and margins of error appear:
All neutrals shifted to: |
New Percentage |
Margin of Error |
Positive |
42.50 |
28.85 |
Negative |
72.50 |
11.54 |
These numbers indicate a peculiar affinity for negative connotations with the “get +” construction, at least in the UKSPOK corpus.
It should be noted that the first construction to appear in the t-score listings is almost invariably a positive one, which makes an interesting exception to the negative-connotation trend. The construction is “get@ + married.” The t-scores for this construction are as follows:
Tense |
t-score (all corpora) |
get married |
|
getting married |
14.320537 |
got married |
19.616369 |
|
Ave: 18.037452 |
An examination of actual examples of all constructions in the Cobuild corpus, however, confirms the negative-connotation tendency. Since actual pages of data are somewhat less than useful, the reader is advised to enter the following queries into either the UKSPOK or all corpora (UKSPOK will give a smaller set of examples, which doesn’t say much since they number in the low thousands):
get@+VERB
get@+JJ
get@+IN
The last query has been offered so that the reader may verify the impracticality of an analysis of the prepositional “get +” construction.
Overall, it appears that the questions asked at the beginning of this analysis have been answered, at least preliminarily. A further inquiry into different negative connotations of verbs and adjectives might also prove instructive, to provide a more definite structure of usage for the “get +” construction. A next-generation build of grammatical search software would also be useful, to attempt the analysis of the prepositional form, and see if the negative-connotation trend holds there as well.
Appendix 1: “get” construction t-scores, sifted & sorted
"get"
t-score |
0:1
UKSPOK |
|
|
|
|
|
|
. - good |
|
|
|
/ - bad |
|
|
|
` -
neutral |
|
|
|
|
|
|
|
WORD |
FREQUENCY |
OCCURRENCE |
T-SCORE |
|
|
|
|
.better |
3604 |
72 |
7.392041 |
.elected |
103 |
6 |
2.341257 |
.engaged |
86 |
7 |
2.562086 |
.excited |
124 |
12 |
3.371966 |
.extra |
626 |
10 |
2.652745 |
.fed |
262 |
20 |
4.321342 |
.free |
816 |
9 |
2.29989 |
.involved |
1761 |
219 |
14.492357 |
.married |
968 |
166 |
12.690716 |
.offered |
337 |
7 |
2.317899 |
.paid |
983 |
81 |
8.718869 |
.pregnant |
211 |
24 |
4.78812 |
.promoted |
44 |
6 |
2.403254 |
.saved |
129 |
6 |
2.313936 |
.sorted |
254 |
7 |
2.398646 |
.straight |
1056 |
10 |
2.302747 |
.sucked |
31 |
5 |
2.200384 |
.together |
2477 |
127 |
10.703682 |
.treated |
261 |
7 |
2.391836 |
/
addicted |
80 |
15 |
3.819816 |
/ angry |
217 |
16 |
3.860364 |
/ annoyed |
127 |
15 |
3.788581 |
/
arrested |
57 |
8 |
2.776556 |
/ beaten |
95 |
7 |
2.55333 |
/ blamed |
30 |
7 |
2.616566 |
/ bored |
203 |
39 |
6.16133 |
/ caught |
367 |
23 |
4.598862 |
/ changed |
1253 |
19 |
3.619001 |
/ charged |
101 |
6 |
2.343359 |
/ cold |
584 |
12 |
3.030172 |
/
confused |
117 |
12 |
3.377167 |
/ cross |
571 |
24 |
4.598975 |
/
depressed |
119 |
7 |
2.529982 |
/ drunk |
160 |
16 |
3.897043 |
/
expelled |
28 |
6 |
2.420067 |
/ fat |
220 |
7 |
2.431723 |
/
frustrated |
72 |
9 |
2.938226 |
/ hit |
604 |
14 |
3.326158 |
/ hooked |
50 |
12 |
3.42695 |
/ hurt |
315 |
11 |
3.072163 |
/ ill |
396 |
7 |
2.260501 |
/ killed |
269 |
13 |
3.413517 |
/ locked |
151 |
6 |
2.290818 |
/ lost |
884 |
29 |
4.962642 |
/ mugged |
22 |
11 |
3.299551 |
/ nervous |
165 |
6 |
2.276107 |
/ older |
1155 |
64 |
7.628388 |
/ shot |
210 |
8 |
2.637322 |
/ sick |
276 |
8 |
2.577261 |
/
stressed |
141 |
23 |
4.720157 |
/ stuck |
487 |
30 |
5.248368 |
/ taken |
1444 |
14 |
2.748312 |
/ teased |
12 |
5 |
2.222255 |
/ thrown |
194 |
14 |
3.608202 |
/ tired |
218 |
25 |
4.887777 |
/ upset |
318 |
10 |
2.903442 |
/ used |
8247 |
111 |
8.520854 |
/ weighed |
29 |
5 |
2.202686 |
/ worried |
503 |
11 |
2.926262 |
/ worse |
687 |
25 |
4.646342 |
` behind |
854 |
14 |
3.15418 |
` dressed |
141 |
25 |
4.927415 |
` enough |
2703 |
57 |
6.628312 |
` mixed |
362 |
15 |
3.632403 |
` more |
16107 |
190 |
10.776344 |
` over |
7618 |
115 |
8.895329 |
` picked |
508 |
11 |
2.922381 |
` ready |
604 |
24 |
4.581637 |
` sent |
718 |
9 |
2.383972 |
` smaller |
383 |
9 |
2.671395 |
` started |
2882 |
17 |
2.323958 |
` told |
2454 |
24 |
3.609643 |
Appendix 2: “getting” construction t-scores, sifted & sorted
"getting" |
t-score
0:1 UKSPOK |
|
|
|
|
|
|
. - good |
|
|
|
/ - bad |
|
|
|
` -
neutral |
|
|
|
|
|
|
|
WORD |
FREQUENCY |
OCCURRENCE |
T-SCORE |
|
|
|
|
.better |
3604 |
61 |
7.595665 |
.bigger |
471 |
15 |
3.816431 |
.excited |
124 |
6 |
2.425949 |
.married |
968 |
53 |
7.218278 |
.paid |
983 |
20 |
4.36992 |
.pregnant |
211 |
8 |
2.793736 |
.ready |
604 |
19 |
4.294461 |
.stressed |
141 |
10 |
3.141543 |
.together |
2477 |
13 |
3.286079 |
/bored |
203 |
4 |
1.9528 |
/caught |
367 |
8 |
2.768088 |
/confused |
117 |
8 |
2.809191 |
/drunk |
160 |
7 |
2.617629 |
/killed |
269 |
4 |
1.937454 |
/nowhere |
194 |
11 |
3.289424 |
/older |
1155 |
19 |
4.235678 |
/paranoid |
36 |
4 |
1.99163 |
/smaller |
383 |
5 |
2.156417 |
/stuck |
487 |
6 |
2.357034 |
/tired |
218 |
4 |
1.949312 |
/worse |
687 |
49 |
6.954361 |
`all |
34233 |
39 |
3.695875 |
`dark |
319 |
7 |
2.589683 |
`dressed |
141 |
6 |
2.422721 |
`fed |
262 |
11 |
3.27989 |
`involved |
1761 |
36 |
5.863515 |
`mixed |
362 |
7 |
2.582125 |
`pretty |
1292 |
5 |
1.967375 |
`used |
8247 |
23 |
3.996162 |
`wet |
246 |
6 |
2.402788 |
Appendix 3: “got” construction t-scores, sifted & sorted
"got"
t-score |
0:1
UKSPOK |
|
|
|
|
|
|
. - good |
|
|
|
/ - bad |
|
|
|
` -
neutral |
|
|
|
|
|
|
|
WORD |
FREQUENCY |
OCCURRENCE |
T-SCORE |
|
|
|
|
.better |
3604 |
39 |
4.340158 |
.bigger |
471 |
10 |
2.670661 |
.engaged |
86 |
11 |
3.231038 |
.fed |
262 |
30 |
5.319338 |
.invited |
162 |
6 |
2.231194 |
.married |
968 |
251 |
15.641308 |
.paid |
983 |
14 |
2.874505 |
.plenty |
323 |
35 |
5.735872 |
.pregnant |
211 |
18 |
4.078486 |
.promoted |
44 |
8 |
2.77708 |
.saved |
129 |
7 |
2.484818 |
/angry |
217 |
6 |
2.157081 |
/arrested |
57 |
5 |
2.151929 |
/attacked |
99 |
5 |
2.089932 |
/bored |
203 |
15 |
3.699979 |
/broke |
272 |
7 |
2.306418 |
/broken |
368 |
8 |
2.398981 |
/caught |
367 |
30 |
5.256063 |
/chucked |
51 |
6 |
2.380767 |
/cross |
571 |
8 |
2.162085 |
/divorced |
70 |
12 |
3.397404 |
/drunk |
160 |
7 |
2.446144 |
/fined |
20 |
5 |
2.206546 |
/hit |
604 |
8 |
2.123575 |
/hurt |
315 |
6 |
2.025026 |
/kicked |
111 |
6 |
2.299917 |
/killed |
269 |
13 |
3.359295 |
/lost |
884 |
19 |
3.689505 |
/nowhere |
194 |
28 |
5.170491 |
/stabbed |
17 |
9 |
2.981296 |
/stolen |
107 |
5 |
2.078123 |
/stuck |
487 |
17 |
3.733244 |
/tired |
218 |
7 |
2.373786 |
/worse |
687 |
33 |
5.349827 |
`different |
6298 |
41 |
3.156614 |
`involved |
1761 |
63 |
7.204944 |
`mixed |
362 |
10 |
2.784432 |
`older |
1155 |
43 |
5.976067 |
`together |
2477 |
49 |
5.832024 |
`used |
8247 |
76 |
5.59535 |