[an error occurred while processing this directive]
The word guanxi (collocation) and meanings of bribe: Deeply rooted, disgusting, sad endings
What makes bribe so deeply-rooted and disgusting? While this
may be very difficult, if not impossible, to quantify, it may be something
we can do rather conveniently from the linguistical point of view using some help from statistics.
Using Hunston's words, "the frequent co-occurrence of a lexical item
with items expressing a particular evaluative meaning" [1a] [1b]
gives the sementic preference or the attitudinal preference of a word,
i.e. the true meaning of a word, or alternatively, the guanxi of word,
or word guanxi (relationship), can be found by understanding the words that appear around it.
So, by looking at the words that co-occur (relate, or having a guanxi) with the word "bribe",
we can probably understand more about its true meaning.
Photo © Alexandra Dubovski
To determine the most collocated (co-occuring) words for "bribe",
let us turn our attention to the Corpus of Historical American English (COHA) ,
an English text database maintained by Brigham Young University, Provo, UT (USA).
The database (or corpus) contains an archive of American text selected from books,
magazines, novels, etc., published within the last 200 years (1800-2009).
By searching through COHA, a total of 226 collocates of "bribe" with a
"Mutual information" (MI) value of no less than 3.0  or
a raw occurrence frequency of at least 1,000 (that is, total occurence count
multiplied by occurence percentage with core word) were retrieved from COHA.
These words are then tabulated and grouped into the following tables.
Each row lists one co-occuring word (or technically known as a collocate),
its total occurence count in COHA, the occurence percentage found with the core word,
and its corresponding MI values.
A further examination of the 226 collocates reviews that approximately 90%
of them fall into several rather distinct groups or meaningful categories.
For simplicity reasons, I shall mainly focus on those with an MI value
greater than or at least very near 5.0 (those highlighted in red).
*** Nature in General ***
CONSPIRACY 5230 0.78 M5.94
CRIMES 7230 0.28 M4.44
GUILTY 17779 0.29 M4.52
ILLEGAL 6037 0.27 M4.37
MONEY 146465 0.03 M3.48
PLOT 8532 0.15 M3.58
THEFT 2618 0.31 M4.58
UNLAWFUL 1473 0.27 M4.41
VIOLATIONS 2218 0.14 M3.40
*** Ease and Directness ***
ACCESSIBLE 2932 0.17 M3.74
INDIRECT 2915 0.17 M3.75
*** Possible Style ***
MONTHLY 5432 0.11 M3.11
SYSTEMATIC 3222 0.25 M4.28
*** Size and/or Scale ***
GRANNY 2411 0.12 M3.28
PALTRY 1172 0.34 M4.74
PETTY 5335 0.17 M3.72
LAVISH 2370 0.21 M4.05
COLOSSAL 2113 0.14 M3.47
*** Involvement ***
INVOLVING 6283 0.29 M4.49
PARTICIPATING 1809 0.17 M3.70
RELATING 4739 0.11 M3.05
*** How it started ***
PERVERTED 1284 0.23 M4.19
RESORT 8076 0.11 M3.12
RESORTED 2401 0.25 M4.29
RESORTING 789 0.38 M4.90
PECUNIARY 2244 0.13 M3.39
*** How it proceeds ***
BEGGED 7298 0.11 M3.10
COAXED 1018 0.79 M5.94
COAXING 781 0.26 M4.33
INDUCE 4826 0.25 M4.28
INDUCEMENT 1167 0.34 M4.75
INDUCING 814 0.37 M4.85
INTRIGUE 1689 0.77 M5.91
INTRIGUES 1053 0.76 M5.89
LURED 1508 0.20 M3.96
NEGOTIATION 2509 0.16 M3.64
OVERLOOK 2390 0.17 M3.71
PERSUADE 6358 0.11 M3.11
PERSUASION 2500 0.40 M4.97
POSE 3480 0.14 M3.49
PROCURED 3000 0.17 M3.71
PROFFERED 1351 0.37 M4.86
PROMISES 8720 0.22 M4.09
TEMPT 1998 0.30 M4.56
TEMPTING 2296 0.26 M4.35
OUTRIGHT 2905 0.59 M5.52
UNSCRUPULOUS 1592 0.25 M4.30
Table 1: General nature of "bribe"
General meaning of bribe – my personal interpretation
Table 1 contains collocates of bribe that describe its meaning in general.
Unlike armed robbery or any other violent crimes, bribe, as described by
its collocates, can be (in my opinion) thought of as a slowly proceeding,
gently persuaded, coaxed (MI=5.94) and intrigued (MI=5.89, MI=5.91)
crime that normally disguises itself as an open, direct and outright (MI=5.52)
action involving paltry (MI=4.74) amount of money.
It is, in fact, one that is well-planned, a conspiracy (MI=5.94),
and involves lavish (MI=4.05) amount of money, possibly both in a systematic (MI=4.28)
and even monthly (MI=3.11) manner.
However, does that mean petty corruption is something we do not have to
worry? I'm afraid not. Bearing in mind that our statistics come from COHA,
Corpus of Historical American English, the collocates here reflect only
the American scenario, and obviously not for Nigeria, Pakistan or
any other countries.
[an error occurred while processing this directive]
Note 1a: For explanations on the "lexical item", please see my other article on "John Sinclair's lexical item".
Note 1b: See p. 265-266 in Hunston, S. (2007). Semantic prosody revisited. International Journal of Corpus Linguistics, 12(2), 249-268.
Note 2: Corpus of Historical American English at Brigham Young University, see
Note 3: A Mutual Information (MI) value of 2.0 or 3.0 means high frequency "noise words"
like the, of, is, am, are, etc. will be ignored. Typically, MI values of 3.0 or above
shows a "semantic bonding" between the words and are therefore more meaningful in
terms of collocation analysis. Too small MI values indicate words simply collocates,
for example, grammatical words like is, am, are, etc. and are therefore ignored
in this research. Meanwhile, a major limitation of this research requires that only
content words will be taken as a core whereas other meaningful words, like the phrasal verbs,
are not considered.