Tactful or Doubtful? Expectations of Politeness Explain the Severity Bias in the Interpretation of Probability Phrases

Laboratoire Travail et Cognition, Centre National de la Recherche Scientifique, Toulouse, France.
Psychological Science (Impact Factor: 4.43). 10/2006; 17(9):747-51. DOI: 10.1111/j.1467-9280.2006.01776.x
Source: PubMed
When a statement about the occurrence of a medical condition is qualified by an expression of probability, such as the word possible, listeners interpret the probability of the condition as being higher the more severe the condition. This severity bias can have serious consequences for the well-being of patients. We argue that the bias is due to a misconception of the pragmatic function served by the expression of probability. The more severe the condition, the greater the chance that the listener construes the expression as a politeness marker rather than as an uncertainty marker. When this misconception does not occur, neither should the severity bias. An analysis of interpretations of probability expressions using a membership-function approach validates this account. We discuss the consequences of this bias for the communication of risk within and outside the medical domain.


Available from: Gaelle Vallee-Tourangeau, Jun 14, 2014
Tactful, or Doubtful? Expectations of Politeness explain the
Severity Bias in the Interpretation of Probability Phrases
Jean-François Bonnefon
CNRS, Toulouse
Gaëlle Villejoubert
Leeds University Business School
When a medical condition is qualified by a probability phrase such as “possible,” hear-
ers assign higher numerical probabilities to more severe conditions. This severity bias
can have serious consequences for the well-being of patients. It is argued that the bias
is due to a misconception of the pragmatic function served by the probability phrase.
The more severe the condition, the greater the chance that the phrase is construed as
a politeness marker rather than as an uncertainty marker by the person who hears it.
When this misconception does not occur, neither should the severity bias. An analysis of
hearer’s interpretations of probability phrases using a membership function approach
validates this account. We discuss the consequences of this bias for the communication
of risk within and outside the medical domain.
Natural language is a poor tool when it comes to com-
municating the likelihoods of states of affairs. Probabili-
ties, of course, are a much better medium for communi-
cation of uncertainty, which is why most of us prefer to
be given numerical rather than verbal estimates of like-
lihoods (Wallsten, Budescu, Zwick, & Kemp, 1993). Un-
fortunately, numerical estimates are often unavailable,
forcing us to base our decisions on phrases as ambigu-
ous as X is rather likely” or Y is highly possible”.
Not surprisingly, people are prone to a number of
biases in their interpretation of probability phrases.
Prominent among these is the severity bias, which is
usually demonstrated with health-related material (Fis-
cher & Jungerman, 1996; Franic & Pathak, 2000; We-
ber & Hilton, 1990). When a probability phrase (prob-
able, possible, likely, etc.) qualifies a medical condi-
tion a patient might develop or a potential side effect
of some treatment, it receives a numerical interpreta-
tion that is all the higher for more severe conditions.
The severity bias can have serious consequences for the
well-being of patients, especially as they become more
directly involved in medical decisions. Overestimating
the likelihood of a side effect can encourage the choice
of the wrong treatment. Misunderstanding the likeli-
hood of developing a given condition can make com-
munication between doctor and patient frustrating and
Jean-François Bonnefon, Laboratoire travail et cognition,
Centre national de la recherche scientifique, Toulouse, France.
Gaëlle Villejoubert, Leeds University Business School, the Uni-
versity of Leeds, Uni ted Kingdom. We thank Reid Hastie,
Mansur Lalljee, Linda Moxey, and Karl Teigen.
Send correspondence to Jean-François Bonnefon, Labo-
ratoire travail et cognition, Maison de la recherche, 5 al-
lées A. Machado, 31058 Toulouse Cedex 9, France. E-mail:
counter-productive. This potential for damage is fur-
ther increased by large-scale policies such as the Eu-
ropean recommendation to give verbal estimations of
side effect likelihood (European Commission, 1998)
some experts have asked for the suspension of such poli-
cies until research has yielded further insight into the
interpretation of probability phrases (Berry, Raynor, &
Knapp, 2003). However, only meager understanding
has been achieved so far.
Insight into the mechanics of the severity bias has
gone no further than Weber and Hilton’s initial explana-
tion that “more severe events may draw attention to po-
tentially higher probability levels, something that might
be labeled a ‘worry effect”’ (1990, p. 788). We propose
that the severity bias derives from what Goffman (1967)
and Brown and Levinson (1978/1987) have identified
as a fundamental mechanism of human social life, face-
work. All humans project a sense of positive identity
and public self-esteem called “face”, and all of us are
motivated to support our own and other’s face in social
interactions. Many actions, called face-threatening acts,
can induce a loss of face (e.g., disagreeing, criticizing,
giving orders, embarrassing others). When such an ac-
tion is performed, the actor is likely to resort to one of
many linguistic strategies that mitigate the face threat.
Among these strategies is the use of probability phrases,
not to communicate degrees of uncertainty, but rather
to reduce the impact of the face-threatening acts.
Our explanation of the severity bias involves two
assumptions. First, as an alternative to their stan-
dard function of likelihood communication, probabil-
ity phrases can be used as face-management devices
when they qualify face-threatening acts (e.g., “you
might be misinformed,” “it is possible you will have to
pay for my lunch”). There is ample evidence for this
claim, both qualitative (Brown & Levinson, 1978/1987)
and quantitative. For example, Youmans (2001) re-
Page 1
ports a rate of 787 probability phrases per 20,000
words (from American English speakers), of which 39%
were used for face-management rather than likelihood-
communication purposes. Second, probability phrases
imply high probabilities when they are used for face-
management. The phrase “possibly” in “it is possi-
bly going to snow tomorrow” performs its likelihood-
communication function, and accordingly denotes a
moderate probability of snowfall. Used in “your bad
breath is possibly the reason people shun you,” it has
nothing to do with uncertainty, and all to do with face-
management. In this context, it denotes a high likeli-
hood. Preliminary evidence for this claim is presented
in Bonnefon and Villejoubert (2005) who found that the
terms “possibly” and “probably” denoted higher likeli-
hoods when they qualified criticisms or impositions than
when they qualified non-face-threatening contents.
All other things being equal (e.g., for similar base
rates), probability phrases denote the same probability
whatever the condition they qualify, as long as they are
perceived to perform their likelihood-communication
function. But, when a physician tells a patient she is
going to develop a medical condition, her social “face” is
threatened and the threat is greatest when the condition
is most severe. Therefore, the more severe the patient’s
condition, the more likely a probability phrase will be
interpreted as a face-management device, rather than as
a likelihood-communication device. Increasingly severe
conditions increase the number of speakers and hearers
who switch to a face-management interpretation of the
phrase. This, in turn, increases the average probability
attached to the phrase, hence the severity bias.
Participants were recruited by third-year psychology
students at the University of Toulouse, as a course credit
requirement. Each student made a list of several men
and women, older than 18 and not studying psychol-
ogy. Then each student randomly drew one man and
one woman from his or her list, and asked them to take
part in the study. Of the 810 participants in the final
sample (401 men, 409 women, mean age = 31.2, SD
= 12.8), 21% had completed graduate school, 47% had
an undergraduate education, 19% graduated from high
school only, and the educational level of the remaining
13% was lower than high school. The sample included a
large proportion of students (39%), but 61% came from
a great variety of non-student professions (including 8%
Material and Procedure
Data collection focused on the phrase “possibly”. The
numerical interpretation of this phrase was assessed by
eliciting its fuzzy membership function (Zadeh, 1965).
This membership function assigns a number to each
value on the probability line [0, 1] that represents its
degree of membership in the concept defined by the
phrase. Degree of membership is usually expressed as a
real number from 0 to 1, such that memberships of 0 de-
note probabilities that are absolutely not in the concept,
and memberships of 1 denote elements that are perfect
exemplars of the concept. Other values represent inter-
mediate degrees of membership. The membership func-
tion approach (originally suggested by Wallsten, Bude-
scu, Rapoport, Zwick, & Forsyth, 1986, and Rapoport,
Wallsten, & Cox, 1987) provides a subtle and rich rep-
resentation of the meaning of probability phrases, and
has been carefully validated in many studies (see for re-
views Budescu & Wallsten, 1995 or Karelitz & Budescu,
Membership functions were elicited using the Mul-
tiple Stimuli Method introduced in Budescu, Karelitz,
and Wallsten (2003). Participants were asked to imag-
ine that their family doctor had announced they would
“possibly” develop some condition during the year to
come. One condition was deafness, the other insomnia
(order of presentation was counterbalanced). Insom-
nia and deafness are of similar prevalence in the French
population, from which participants were sampled (i.e.,
with incidence rates of about 4% in year 2000).
After having read, “The doctor tells you, you will
possibly suffer from insomnia soon,” participants were
asked: “Does the doctor think the probability that you
will suffer from insomnia soon is. . . followed by the 10
values 10%, 20%, up to 100%. They provided a judg-
ment for each of the 10 percentages, on a 10-point scale
anchored at “absolutely not” and “absolutely”. The task
was then repeated with judgments of the second med-
ical condition (deafness or insomnia). In addition to
the membership function judgments, participants were
asked to say which of deafness or insomnia was worse
Finally, they were asked whether the doctor was
qualifying deafness as “possible” because (a) she was
not sure it would happen, or rather because (b) she
wished to announce it tactfully. The same question was
asked with respect to insomnia, to assess in each case
which speech function participants believed the proba-
bility phrase was intended to perform.
“Possible deafness” is more probable than “possible
insomnia.” Membership functions were computed by
averaging membership judgments across participants.
Figure 1 (left panel) depicts the function values of “pos-
sible” for insomnia and deafness, regardless of which
speech function participants assigned to the phrase.
Membership functions show the expected severity bias,
Individuals who failed to answer (n = 16) or who judged
insomnia to be worse news than deafness (n = 131) were fil-
tered out, yielding the final sample of 810 participants. This
choice substantially improved the clarity of the analyzes.
Page 2
Table 1
Mean Peaks of the Membership Functions attached to
“Possible” Insomnia and Deafness, as a function of
whether the Phrase is understood as Performing Likeli-
hood Communication or Face Management
M (SD) n
Likelihood Comm. .57 (.33) 669
Face Management .71 (.30) 141
Overall .59 (.33) 810
Likelihood Comm. .55 (.34) 324
Face Management .74 (.28) 486
Overall .67 (.32) 810
as the function for deafness peaks at a higher probabil-
ity than the function for insomnia. The bias is confirmed
by the computation of the two functions’ peaks (see Ta-
ble 1). The peak is computed by averaging for each par-
ticipant the probability values which received the high-
est membership ratings, then averaging across partici-
pants the values so obtained.
The peak of the function
for deafness is higher than for insomnia, t(815) = 7.90,
> .999 (Killeen, 2005), d = .23 (the 95%-confidence
interval for this difference is .06 .09).
One is uncertain about “possible insomnia,” but tact-
ful about “possible deafness.” As shown in the last col-
umn of Table 1, the proportion of participants judg-
ing that the doctor used the term “possible” for face-
management purposes is much greater with respect to
deafness (60%) than insomnia (17.4%, z = 17.1, p
.999, Cohen’s h = .9). One noteworthy result is that no
more than 29 participants judged the doctor was tactful
about insomnia but uncertain about deafness. Thus, less
than 4% of the participants directly contradict our hy-
pothesis that more severe conditions increase the like-
lihood that the probability phrase will be perceived as
The numerical interpretation of “possible” is a func-
tion of its being tactful or uncertain. The top-right panel
of Figure 1 depicts the membership functions of “possi-
ble” for insomnia and for deafness only for participants
who judged that the phrase communicated likelihood.
In contrast, the bottom-right panel depicts the same
functions, but for participants who judged the phrase
to be tactful, serving face-management motives. The
functions within each panel are similar, especially in
the top-right (likelihood-communication) panel, but the
difference between the two panels is dramatic with the
functions in the bottom panel peaking at higher proba-
bilities.With respect to insomnia, the peak of the face-
management membership function is greater than the
peak of the likelihood-communication function(see Ta-
ble 1 for descriptive statistics), t(808) = 4.9, p
> .999,
d = .46 (the 95%-confidence interval for this difference
is .09 .21). Similarly, with respect to deafness, the
peak of the face-management membership function is
greater than that of the likelihood-communication func-
tion, t(808) = 8.5, p
> .999, d = .61 (with a 95%-
confidence interval of .14 .23).
We finally consider the sub-sample of participants
(n = 295) who judged that the doctor was expressing
uncertainty about both insomnia and deafness. If po-
liteness expectations do underly the severity bias, these
participants should not manifest any bias in their nu-
merical interpretations. Indeed, the peaks of the mem-
bership functions for insomnia (M = .57, SD = .35) and
deafness (M = .56, SD = .34) are practically the same
for these participants (t(294) = .8, p
= .55, d = .05,
with a 95%-confidence interval for the difference be-
tween peaks of -.01 +.03. The two peaks are corre-
lated, r = .87.
Previous research has established that probability
phrases not only communicate degrees of uncertainty
but also serve other pragmatic functions, such as ex-
pressing perspective and drawing attention to the occur-
rence or the non-occurrence of the event to which they
refer (Sanford & Moxey, 2003; Teigen & Brun, 2003)
or social face-management. We conjectured that the so-
called severity bias in interpretations of verbal probabil-
ity phrases results from the use of verbal phrases as face-
management devices whose function is to safeguard the
hearer’s feelings while receiving face-threatening news.
We tested this conjecture in a medical context where
a doctor tells a patient that she might develop a
mild or severe medical condition. We expected the
face-management interpretation would be more likely
when the condition was more severe, and that a face-
management interpretation would lead to an overes-
timation of the probability of this condition. Results
replicated the severity bias effect. The same probability
word (“possibly”) was judged to communicate higher
numerical probabilities when it qualified a more severe
condition (deafness) than when it qualified a less se-
vere but equally prevalent condition (insomnia). Fur-
thermore, when the verbal probability phrase qualified
the more severe condition, most participants thought
it served a face-management purpose. Those who
believed the probability phrase was used as a face-
management device also thought that the condition it
qualified was substantially more likely to occur, com-
pared to those who thought the phrase was communi-
cating a vague likelihood. But, those who believed the
probability phrase was intended to communicate the
likelihoods of occurrence for both diseases did not ex-
hibit the severity bias.
Some small discrepancy is common between the numeri-
cal computation of the peak and its graphical representation:
Accordingly, the values in Table 1 do not exactly match the
peaks of the functions in Figure 1.
Page 3
Figure 1. Membership functions of “possible” deafness and insomnia, for the whole sample (left panel, n = 810), for those par-
ticipants who understood “possible” as serving likelihood-communication purposes (top-right panel, n = 669 for insomnia and
n = 324 for deafness), and for those participants who understood “possible” as serving face-management purposes (bottom-right
panel, n = 141 for insomnia and n = 486 for deafness)
.1 .2 .3 .4 .5 .6 .7 .8 .9 1
Thus, people recognize that the more severe a condi-
tion is, the more threatening the news telling someone
they have this condition will be. They also understand
that it is polite and tactful to mitigate face-threatening
news by using linguistic moderators, such as probabil-
ity phrases. Finally, they recognize that a probability
phrase used as a face-management device does not refer
to the probability of the event it qualifies.
While our experimental test of this account was lim-
ited to the phrase “possible,” it should apply to other
phrases as well. However, some phrases may be less
appropriate than others to express politeness. For ex-
ample, we suspect that “probable” is a less plausible po-
liteness term than “possible.” If so, fewer respondents
would make a tactful interpretation of information com-
municated with “probable” as a qualifier. But, this
would not undermine for our main findings. Although
a phrase with the term “probable” in it is less likely to
be interpreted as tactful face-management, when it is
interpreted as “tactful,” it would still be judged to be
a higher numerical probability, than when interpreted
as communicating likelihood information directly (Bon-
nefon & Villejoubert, 2005). Moreover, an uncertain
“probable” should receive the same numerical interpre-
tation whatever the severity of the condition to which it
A misunderstanding about which function a probabil-
ity phrase is intended to serve could lead to a discrep-
ancy between the level of probability a doctor intends
to communicate and the level of probability understood
by her patient. This miscommunication can give rise
to an overestimation of the probability communicated
when the patient interprets a particular phrase as a face-
management device whereas the doctor used it to com-
municate a vague likelihood. Thus, this research sug-
gests that measures should be taken to ensure speakers
and hearers assign a likelihood communication function
to a given probability phrase to improve risk communi-
Berry, D., Raynor, D., & Knapp, P. (2003). Communicating
risk of medication side effects: an empirical evaluation
of EU recommended terminology. Psychology, Health and
Medicine, 8, 251–263.
Bonnefon, J. F., & Villejoubert, G. (2005). Communicating
likelihood and managing face: Can we say it is probable
when we know it to be certain? In B. G. Bara, L. Barsa-
lou, & M. Buciarelli (Eds.), Proceedings of the 27
Conference of the Cognitive Science Society.
Page 4
Brown, P., & Levinson, S. C. (1987). Politeness: Some univer-
sals in language usage. Cambridge: Cambridge University
Press. (Original work published 1978)
Budescu, D. V., Karelitz, T. M., & Wallsten, T. S. (2003). De-
termining the directionality of probability words from their
membership functions. Journal of Behavioral Decision Mak-
ing, 16, 159–180.
Budescu, D. V., & Wallsten, T. S. (1995). Processing linguistic
probabilities: General principles and empirical evidence.
In J. Busemeyer, D. L. Medin, & R. Hastie (Eds.), Decision
making from a cognitive perspective (pp. 275–318). San
Diego, CA: Academic Press.
European Commission. (1998). A guideline on the readabil-
ity of the label and package leaflet of medicinal products for
human use. EC Pharmaceuticals Committee.
Fischer, K., & Jungerman, H. (1996). Rarely occurring
headaches and rarely occuring blindness: is rarely-rarely?
Journal of Behavioral Decision Making, 9, 153–172.
Franic, D. M., & Pathak, D. S. (2000). Communicating the fre-
quency of adverse drug reaction to female patients. Drug
Information Journal, 34, 251–272.
Goffman, E. (1967). Interaction ritual: Essays on face-to-face
behavior. Garden City, NY: Anchor Books.
Karelitz, T. M., & Budescu, D. V. (2004). You say “probable”
and I say “likely”: Improving interpersonal communication
with verbal probability phrases. Journal of Experimental
Psychology: Applied, 10, 25–41.
Killeen, P. R. (2005). An alternative to null-hypothesis signifi-
cance tests. Psychological Science, 16, 345–353.
Rapoport, A., Wallsten, T. S., & Cox, J. A. (1987). Direct
and indirect scaling of membership functions of probability
phrases. Mathematical Modeling, 9, 397–417.
Sanford, A. J., & Moxey, L. M. (2003). New perspectives on
the expression of quantity. Current Directions in Psycholog-
ical Science, 12, 240–243.
Teigen, K. H., & Brun, W. (2003). Verbal expressions of
uncertainty and probability. In D. Hardman & L. Macchi
(Eds.), Thinking: psychological perspectives on reasoning,
judgement, and decision-making (pp. 125–145). Chichester,
UK: Wiley.
Wallsten, T. S., Budescu, D. V., Rapoport, A., Zwick, R., &
Forsyth, B. (1986). Measuring the vague meanings of prob-
ability terms. Journal of Experimental Psychology: General,
115, 348–365.
Wallsten, T. S., Budescu, D. V., Zwick, R., & Kemp, S. M.
(1993). Preference and reasons for communicating proba-
bilistic information in numerical or verbal terms. Bulletin
of the Psychonomic Society, 31, 135–138.
Weber, E. U., & Hilton, D. J. (1990). Contextual effects in
the interpretation of probability words: perceived base-rate
and severity of events. Journal of Experimental Psychology:
Human Perception and Performance, 16, 781–789.
Youmans, M. (2001). Cross-cultural differences in polite epis-
temic modal use in American English. Journal of Multilin-
gual and Multicultural Development, 22, 57–73.
Zadeh, L. A. (1965). Fuzzy sets. Information and Control, 8,
Page 5
  • Source
    • "The explicit demand of evaluating a speaker's feeling of knowing throughout our task could also limit the generalizability of these findings to everyday social situations in which expressions of confidence are irrelevant to the topic but unconsciously attended. For instance, we implicitly process confident vocal cues when others are trying to persuade us (e.g., in marketing behavior, Smith, De Houwer, & Nosek, 2013) and detect vocal cues of uncertainty when a speaker attempts to mitigate a face-threatening situation to the listener (e.g., in doctor-patient communication, Bonnefon & Villejoubert, 2006). Research on emotion shows that effects of integrating and continuously evaluating verbal and vocal information in conflicting statements are greater when vocal cues to the intended speaker meaning are task-relevant, yielding more extensive ERP effects when vocal cues are explicitly attended to (producing a larger N400 effect in the explicit than the implicit condition but a late positivity only in the explicit condition, see Kotz & Paulmann, 2007 for details). "
    [Show abstract] [Hide abstract] ABSTRACT: Listeners often encounter conflicting verbal and vocal cues about the speaker's feeling of knowing; these "mixed messages" can reflect online shifts in one's mental state as they utter a statement, or serve different social-pragmatic goals of the speaker. Using a cross-splicing paradigm, we investigated how conflicting cues about a speaker's feeling of (un)knowing change one's perception. Listeners rated the confidence of speakers of utterances containing an initial verbal phrase congruent or incongruent with vocal cues in a subsequent statement, while their brain potentials were tracked. Different forms of conflicts modulated the perceived confidence of the speaker, the extent to which was stronger for female listeners. A confident phrase followed by an unconfident voice enlarged an anteriorly maximized negativity for female listeners and late positivity for male listeners, suggesting that mental representations of another's feeling of knowing in face of this conflict were hampered by increased demands of integration for females and increased demands on updating for males. An unconfident phrase followed by a confident voice elicited a delayed sustained positivity (from 900 ms) in female participants only, suggesting females generated inferences to moderate the conflicting message about speaker knowledge. We highlight ways that verbal and vocal cues are real-time integrated to access a speaker's feeling of (un)knowing, while arguing that females are more sensitive to the social relevance of conflicting speaker cues. (PsycINFO Database Record
    Full-text · Article · Apr 2016 · Journal of Experimental Psychology Human Perception & Performance
  • Source
    • "A probable explanation is that a few cannot be used to denote zero or all, and is thus unsuitable for hedging statements of the form all people X-ed, or no one X-ed. The results obtained in this study contribute to the growing evidence that quantifiers and verbal probability phrases play an important role in everyday communication as expressions of politeness (Bonnefon et al., 2009; Bonnefon, Fenney, & De Neys, 2011; Bonnefon & Villejoubert, 2006). Face-threatening acts may be modified with those expressions in order to soften their potential effects on the hearer. "
    [Show abstract] [Hide abstract] ABSTRACT: Politeness theory defines face as a sense of positive identity or desired public image that people are motivated to maintain during social interaction. Many actions or utterances are possible threats to the face, so speakers often use linguistic politeness strategies that mitigate these threats. One of politeness strategies is hedging, or weakening a statement that a speaker knows to be true and which could present a threat to the face of the hearer. Bonnefon, Feeney, & Villejoubert (2009) showed that in face-threatening contexts expression some X-ed may be interpreted as all X-ed. The study aimed to investigate the interpretation of quantifiers some, few, and a few in face-threatening contexts. We expected that interpretation of few would differ from interpretations of some and a few due to the property of focus. One hundred and seventy-one participants read three scenarios and judged the appropriateness of quantifier use in different contexts. We replicated the results reported by Bonnefon et al. (2009): participants were more likely to endorse the inference from some X-ed to all X-ed in a face-threatening context than in a face-boost context. Similarly, few X-ed was more likely to be interpreted as no one X-ed in a face-threatening context. For a few there were no differences between the two contexts.
    Full-text · Article · Jan 2014
  • Source
    • "In a face-to-face setting where the subject (i.e., James) is also present, politeness concerns will usually lead to an even greater reluctance to express negative information. Empirical studies have shown that people are sensitive to pragmatic implications of these politeness concerns and adjust their understanding accordingly (e.g., Bonnefon, Feeney, & Villejoubert, 2009; Bonnefon & Villejoubert, 2006; Demeure, Bonnefon, & Raufaste, 2008, 2009 Juanchich, Sirota, & Butler, 2012;). The Bayesian framework also allows clear predictions about the impact of whatever positive information is provided in addition to that which is perceived to be missing. "
    [Show abstract] [Hide abstract] ABSTRACT: “Damned by faint praise” is the phenomenon whereby weak positive information leads to a negative change in belief. This seemingly conflicts with normative Bayesian predictions, which prescribe that positive information should only exert a positive change in belief. We argue that the negative belief change is due to an inference from critical missing evidence; that is, an implicit argument from ignorance. Such an inference is readily incorporated within a version of Bayes’ theorem incorporating the concept of epistemic closure. This reformalisation provides a general theoretical framework for the phenomenon that clearly outlines those conditions under which it should be observed, and its conceptual relationship with other argumentation phenomena.
    Full-text · Article · Jul 2013 · Thinking and Reasoning
Show more