Professional Documents
Culture Documents
3
HOW TO MAKE TEACHING
MATERIALS AND TESTS
FOR PRAGMATICS
In this chapter, I will look at different instruments that can be used to create
activities for teaching and testing L2 pragmatics. Most of these instruments were
developed for pragmatics research, but they can be used just as well for teaching
32
and testing purposes. Two broad types of tools exist: receptive tools, which check
learners’ understanding of pragmatic meaning, and productive tools, which stimu-
late learners to practice the creation of pragmatic meaning. Receptive tools
include metapragmatic judgment and multiple-choice tasks, whereas productive
tools include discourse completion tasks, role plays, and elicited conversations, as
shown in Figure 3.1.
This chapter will focus on how to design these instruments and what kind of
information they can provide to teachers and testers. In the following chapters,
I will draw on these instruments for teaching and testing learners at different pro-
ficiency levels.
The physical context includes the location where the communication happens
(an office, a hospital ward, a living room), the mode of communication (face to
face, over the phone, online, email), and, where relevant, the time: knocking on a
friend’s door and asking them for help at 2 am requires different pragmatics than
doing so at 2 pm.
The social context needs to make clear the social roles of the interlocutors, 34
e.g., student and professor, housemate and housemate, father and son. It also
needs to make clear the settings for the social context variables in Brown and
Levinson’s (1987) framework: Power, (Social) Distance, and Imposition. Power is
usually implied in the social relationship, e.g., a manager is higher in power than
an employee, and a teacher is higher in power than a student whereas housemates,
friends, or colleagues are equal in power.
By contrast, Social Distance is not automatically clear from people’s respective
social roles, and needs to be described in the task. For example, strangers and other
people who have never met before have high social distance whereas people who
have a high degree of social contact with each other (friends, housemates, family
members) have low social distance. People who do not know each other well but
have something in common (a new co-worker, a classmate you have not spoken
to before, a second cousin you have only seen a couple of times) have medium
social distance.
Degree of imposition is a very important driver of the politeness level of the
utterance and must be clear from the situation. As outlined in Chapter 2, high
imposition means high “cost” to the hearer in terms of money, energy, time, or
social reputation. To put it generally, the more they have to go out of their way
to do things they would not ordinarily do, the greater the imposition. Borrowing
your friend’s laptop for an hour is a mid-level imposition: they do not have to do
very much to comply with the request, but they cannot use their laptop during
that time, so there is some cost, but not too much. This situation becomes high
35
imposition if they also need their laptop to finish an urgent task. It becomes low
imposition if they do not need the computer at the moment, or they have a
backup. Putting this information in the task prompt allows the teacher, tester, or
researcher to influence the imposition level.
Especially in testing and research settings, it is very important to make sure
that test takers or research participants perceive the social context settings in the
same way as the tester or researcher. Cultural differences in perceptions can lead
to distortions, so there should always be a piloting with members of the target
community.
Finally, the goal of the interaction needs to be spelled out. What is the speaker
trying to achieve? In real- world talk, people know what they want, but in
pragmatics tasks, we need to tell learners/participants what the point of the talk is.
Including these three major considerations may sound like the task will end
up being very lengthy but the challenge is to keep it concise. That is actually not
as hard as it sounds because not everything needs to be spelled out in great detail.
The following is a scenario for a discourse completion task, designed to elicit a
request:
The physical context is implied as being the workplace setting (restaurant) because
Eric and Nikita are both at work. Since they are sitting next to each other, the
mode of interaction is face-to-face talk.The time is sometime in the morning but
this is not crucial for this situation.
The social relationship between Eric and Nikita is given as colleagues, which
implies low social distance and no power difference. The degree of imposition is
probably medium since Nikita will have to change whatever plans she might have
to accept Eric’s request (which is a cost to her) but at the same time she gets the
benefit of working an extra shift (extra pay).
The goal of the interaction is clearly specified as Eric wanting Nikita to take
his shift, and the scenario also provides a reason for his request. This is important
because in the real world, people do not make requests without a reason. Not
including a reason in the scenario would mean that respondents need to make it
up, which would put extra pressure on them and is not realistic.
As a general guideline, the description of the communicative situation should
be about the length of one paragraph. Anything longer than a paragraph runs the
risk of becoming a reading comprehension test, and respondents are just not likely
36
36 Eric’s request is overly direct and lacks a reason so it would probably be judged
to be mostly inappropriate or completely inappropriate. This illustrates that
metapragmatic judgment tasks do not necessarily have one correct answer: most
respondents would probably agree that Eric’s utterance is not polite enough but
they might disagree as to whether it is “completely inappropriate” or “mostly
inappropriate,” and there is no objective way to decide who is right.
Metapragmatic judgments can also be designed to be overly polite and
have a more specific rating scale that incorporates over-politeness, like the
following task:
Carl’s utterance is much too polite for the situation since the degree of imposition
is quite small; after all, it is the server’s job to sell muffins. Interestingly, overpolite
items were quite difficult in the metapragmatic judgment section of Roever et al.
(2014a) so over-politeness seems to be harder to detect for learners.
A variation on metapragmatic judgment is to ask learners to correct the
inappropriate utterance, which adds a productive element to the task. The
following example shows a metapragmatic judgment similar to the ones used by
Roever et al. (2014a).
If test takers choose “yes,” the test moves to the next item. If they choose “no,”
there are two follow-up questions:
Follow-up questions
If you think it is not appropriate, please explain why.
A. Sarah should have been more polite.
B. Sarah should not refuse a colleague’s request.
C. Sarah’s reason is not good enough to refuse.
D. Sarah should have been more direct.
38
How certain of your response are?
Very certain
Quite certain
Not certain
This design avoids learners having to produce utterances, which might be difficult
because of proficiency limitations, but it means more work for the designer in
creating the response options.
Metapragmatic judgment tasks can also include several target utterances situated
in one scenario, which reduces the need to create a large number of scenarios.
Here is an example:
Appropriate Inappropriate
1. “Oops, my bad.” 5 4 3 2 1
2. “Oh no, I’m so sorry.” 5 4 3 2 1
3. “Sorry, Sally.” 5 4 3 2 1
40
The variations on metapragmatic judgment tasks shown above are only some pos-
sible instantiations. Surprisingly few complete instruments are available. Bardovi-
Harlig and Dörnyei (1998) have made their entire video-based instrument available
on IRIS (www.iris-database.org/). The testing instrument used by Hudson
et al. (1995) is fully reproduced in the appendix of their book. Takimoto (2009)
shows examples of different kinds of metapragmatic judgment tasks, with add-
itional samples available on IRIS. Li (2014) shows some samples of metapragmatic
judgments in L2 Chinese pragmatics.
Possibly surprisingly, there are also no extensive theoretical discussions of
metapragmatic judgments. Culpeper et al. (2018) devote just a subchapter to
metapragmatic judgments, and Taguchi and Roever (2017) even less.
The scenario description for routines items can be kept fairly short with no great
elaboration of the social relationship, since the point of the item is to trigger the
situationally typical formula from a limited range of possibilities, rather than care-
fully craft a highly recipient-designed response.
The challenge with any multiple-choice item is the creation of plausible but
clearly incorrect distractors in addition to the correct option. For routine formulae,
a good way to create distractors is to use formulae that exist but do not fit the
situation. In the example above, the first response option is a correct response to
an apology but the three distractors are responses to expressions of gratitude. This
42
item was medium difficulty in Roever (2005) with 57% of test takers answering
it correctly.
As with all multiple-choice items, no one response option should stick out,
e.g., by being much longer than the others. In the apology item above, all response
options are approximately the same length. Another way to make response options
similar is to start them all off the same way as in the following example:
This item was somewhat difficult in Roever’s (2005) test with 39% of test takers
answering it correctly. Unlike the subway item, the answer choices for the watch
item are not responses to an initiating utterance but they are utterances that open
a conversation. This is demonstrated by each of them starting with “Excuse me,”
which also serves to make them look more similar. 42
Another interesting feature of this item is the second distractor, which was
designed to be attractive to L1 German respondents (who made up a substantial
subset of Roever’s test-taker sample) since it is a direct translation of the equiva-
lent German routine “Wie spät ist es?” Using direct translations is a useful way
to create attractive distractors but can of course only be done in foreign language
settings where all respondents share the same L1.
It is notable that the other distractors are not derived from existing routine for-
mulae but were created for this item. This is not ideal but using routine formulae
to make distractors is not always feasible.
The implied meaning in the above example is that Frank is probably in his room,
making the third answer choice the correct answer. The type of implicature in
the example above is idiosyncratic implicature (Bouton, 1988, 1999) or non-
conventional implicature (Taguchi, 2009, 2011), which does not follow a par-
ticular fixed pattern. This was an easy item on Roever’s (2005) test, which 79% of
test takers answered correctly.
Taguchi (2008, 2009) suggests the category of conventional implicature, which
includes routine formulae as well as commonly used sequential means for implying
pragmatic meaning. Here is an example targeting implied refusal:
42
Implicature item for implied refusal
Mark and Sally are students who share a house. Sally has finished her uni-
versity work in the morning, and has nothing else to do. They are talking
in the living room.
Sally:“Do you want to do something this afternoon? Catch a movie maybe?”
Mark: “Well, I’ve got a paper due tomorrow.”
Sally: “Come on, it’s just a couple of hours.”
Mark: “I really need to get this paper done today.”
What does Mark mean?
1. He does not want to go to a movie today.
2. He will be finished with his paper in about two hours.
3. He wants to do something with Sally in the afternoon.
4. He wants Sally to help him with his paper.
The item above was notably more difficult for learners than the idiosyncratic
implicature item (see the first example in section 3.3.2). Only 31% identified
the first response option as the correct answer. It is a general finding that implicature
based on irony or sarcasm is more difficult for learners, possibly because it has the
extra cognitive step of recognizing that the implied meaning is the opposite of the
surface meaning.
Just like with routine formulae, the creation of distractors is difficult for impli-
cature items.Taguchi (2005) suggests three principles of distractor design: one dis-
tractor should represent the opposite of the correct answer, one distractor should 4
relate to the final utterance, and one should relate to the dialogue as a whole.
Roever constructed his distractors along these lines for his formulaic implicature
item shown above: the first one is the correct answer, the second one relates to the
final utterance, and the third one is the opposite of the correct answer. Only the
fourth one does not follow Taguchi’s principle and is designed as the interlocutors
misunderstanding each other or being deliberately obfuscating.
item with the first answer choice as the intended correct answer and the others
as distractors.
The first response option seems completely fine, including a reason and an
indirect request with a modal and interrogative. The second response looks
much less acceptable as it is a bit too much of a command given the equal power
relationship between Eric and Nikita, and it is lacking an explanation. However,
the directness of it can also be read as expressing great urgency, or maybe Eric
4 is just given to a bit of drama. The missing explanation could be taken care of
by Nikita asking “why, what’s up?” in the next turn, thereby giving Eric space
to explain.
Response option 3 is even more borderline: it seems to ignore the imposition
on Nikita of having to change her plans this afternoon and it actually makes it
sound like taking Eric’s shift is in Nikita’s interest. There is no explanation of why
Eric wants Nikita to take his shift, but since Eric casts the request as an offer, is a
reason really required? Some respondents might feel that it is not polite enough
in disregarding the imposition and not giving a reason but others might find it
unproblematic as casual talk between colleagues. Also, if Nikita really wanted to
know, she might ask in the next turn.
Response option 4 goes into the direction of over- politeness, and some
respondents might feel that it is too formal given Eric’s and Nikita’s relationship.
But it is not grossly inappropriate, and if Eric and Nikita have larger social dis-
tance because they do not know each other well, or there is a significant age gap
between them, a more polite utterance may be justified.
Of course, it would be possible to create response options that are clearly too
direct or too arcane (“Nikita, take my shift this afternoon.” /“Nikita, are you
doing anything this afternoon?”) but they would be too obviously off the mark,
just like a vastly overpolite option (“Nikita, I’m so sorry to bother you, but I was
just wondering if it would be at all possible for you to do me a huge favor and
take my shift this afternoon?”). Overall, it is best not to use multiple-choice tasks
for speech acts.
46
Options 1, 3, 5, and 7 are all likely in the scenario, and an item like this encourages
learners to compare several routines and their relationship to a scenario.
A similar item type could be used for implicature, as shown in the following
example: