blog

In Anita Pomerantz’s canonical paper on preference for agreement and disagreement with assessments in conversation I found two fascinating examples of precisely the phenomena I was looking for in the ways people talk about art.

Pomerantz uses the example below to demonstrate what she calls ‘second-assessment
productions: agreement dispreferred’, by which she means that when someone
produces a self-deprecating assessment in conversation, it invites an agreement
or a disagreement with that self-deprecation – but generally a disagreement is
preferred.

As is often the case with this kind of conversational analysis, the evidence
for one action or response being preferred is found by observing what happens
when the ‘dispreferred’ action or response is supplied. In Pomerantz’s
examples of assessments, ‘dispreferred’ conversational responses, often
contradictions and disagreements, are characterised by delays, pauses, and
other avoidances or ‘softenings’ of the dispreferred response. It is in this
context that Pomerantz produces the example below, as evidence for structure of
how conversational participants tend to withhold what she calls ‘coparticipant
criticism’ – basically, worming their way around insulting each other’s
sensibilities.

However, this extract also shows some of the features of aesthetic
conversations that I intuitively shaped into my research question about how the
criteria for an assessment become relevant referents for sub-asessements in an
aesthetic evaluation.

Pomerantz identifies this entire exchange only as a series of deferrment turns,
in which D’s (dispreferred) critical assessment is delayed, softened or
otherwise minimised.

With my question about the criteria for judgement being negotiated in mind,
these exchanges seem to support the idea that a feature of these deferrments of
dispreferred responses is that they include a negotiation of the referents of
the assessments they precede:

In the exchange highlighted above, A offers up a print for assessment, D gives
a positive assessment (softening an immanent critical assessment, Pomerantz
suggests), and C interrupts, identifying one criterion for judgement: the
question of authorship. A, C and D then negotiate claims to knowledge of this
referent (the author of the print).

Two further criteria are then raised by A – the assessment that the print is
monetarily valuable, which is met with a silence (which Pomerantz sees as an
implicit marker for a dispreferred second assessment – in this case, a
disagreement). A then raises the rarity of the print “only a hundred of’m”,
which D acknowledges after with ‘Hrm’ after a further pause.

Pomerantz observes in this paper that acknowledgements of prior assessments do
not imply claims of access to a referent: by acknowledging A’s claim of the
rarity and value of the piece without agreeing or disagreeing, D acknowledges
only the claim, but (marked by a further pause) does not participate in
claiming to have that knowledge.

E then requests a clarification of the referent being assessed: “Which picture
is that.”, then interrupts D’s negative assessment, seemingly to raise a
further criterion for assessment of the print: the spelling of the word ‘Life’
in the print that A points out. E then seems to claim that “That’s all I wd
loo(hh)k fo(h)” in the print.

Finally, D delivers a critical assessment, raising several further criteria:

• that this print belongs to a type of art that can be more or less
  ‘realistic’, and that D assesses this to be ‘less realistic’. This criterion is coupled with an assessment that D likes the ‘more realistic’ of this ‘type of art’
• That the print belongs to a “magazine advertisement” type, which can be
  more or less “great”, and D implys an assessment that this print is less great.

I am provisionally thinking of aesthetic judgements as assessment sequences in
cases where the referent of the assessment is evidently ambiguous.

Although Pomerantz is only concerned with the overall structures of preference
in agreement and disagreement with assessments, this extract does seem to bear
out some of the assumptions in my research question: namely that in aesthetic
judgements, conversational participants seem to offer up candidates for
assessment criteria along with their assessments.

Other examples of assessment sequences in Pomerantz’s text that deal with more
self-evident referents do not seem to exhibit this characteristic, suggesting
that it may be useful to look at aesthetic judgements as a special case of
conversational interaction.

Another assumption in my research question is that as conversational
participants offer up candidate criteria as referents for their assessments,
the opportunity for ‘topical drift’ is extended. The discussion of one referent
may lead to another, and yet another, potentially replacing the subject of an
initial assessment with a sequence of second and third assessments of different
referents altogether.

In a note on a section about ‘upgrades’ (described in her paper as strong
agreements with assessments on sequential grounds), Pomerantz picks out an
exception to the upgrades she finds in the corpus that reinforce prior
assessments of the same referent: upgrades that also slightly modify the
referent, and then reinforce it:

In this extract A assesses as ‘nice’ the way two things appear together. B
upgrades the assessment to ‘lovely’, but generalises the referent to the two
things – not their appearance together.

Pomerantz identifies this topic shift as part of a softening of a later
dispreferred disagreement with an assessment: B eventually emphasises the
niceness being that the two pieces are separate – contradicting A’s initial
assesment.

Once again, in the interim between assessment and dispreferred disagreement, B
offers up the colours of the pieces (“blue en grey, en white”) as candidate
criteria for assessment, which A agrees with in this sequence.

Again, this seems to bear out some of the intuitions in my research question –
not only that a component of the conversational pragmatics of aesthetic
judgements is the proffering of multiple criteria for assessment, but also that
in this process, there is an opportunity for a shift in the referent being
assessed.

The qualification for these intitutions that I can take from Pomerantz’s paper
is that this proffering of alternate candidate criteria may be seen as a
specific case of deferring or delaying a dispreferred second assessment in a
sequence.

I’ve started my literature review with Anita Pomerantz’s “Agreeing
and disagreeing with assessments: some features of preferred/dispreferred turn shapes” to try and get a flavour of how ‘judgements of taste’ might look
through the lens of conversation analysis (CA).

Pomerantz begins her paper with the assertion that assessments are a
fundamental conversational feature of participation in an activity. She finds
evidence for this claim in the way that people decline to make assessments of things they haven’t participated in.

In the example below, this claim is demonstrated by B declining to deliver an expected assessment of the “dresses” by claiming lack of access to them: “I haven’t been uh by there-“.

Pomerantz then points out that a statement of having participated in something is somehow incomplete without a correlating assessment.

The above example intuitively bears this out: if L’s completion of this
sequence with “it’s really good” is missed out, the conversation would seem
stilted.

Reading Pomerantz’ detailed, CA account of the coordination of assessment patterns in speech alongside Kant and Hume’s philosophical
discourses on aesthetics and judgement offers some compelling but potentially incompatible insights.

Pomerantz’s observations of how preference structures work in assessments, namely that there are ‘preferred’ and ‘dispreferred’ responses to assessments
seems to map onto the way Kant uses ‘blame’ in judgements of taste.

Kant’s critique of aesthetic judgement asserts that an aesthetic judgement is
distinctive because we ‘blame’ others for not agreeing with us. Kant uses this
as a way of differentiating between judgements of the ‘agreeable’ and
judgements of taste: when something is ‘agreeable’ (such as, for
example, a cute puppy), we don’t argue the point with people who don’t like
dogs because, even when most people seem to find them agreeable, not everyone
has to like all kinds of puppies. However, when discussing aesthetic
judgements, we argue the point. For Kant, that willingness to argue marks out
an aesthetic judgement from other forms of judgement: a judgement that is (at
least potentially, in the mind of the judger) a universal judgement.

Pomerantz bases her claim that there are ‘preferred’ and ‘dispreferred’
responses to assessments on the regularly observable structures of how people
negotiate assessments in everyday conversation.

She identifies preferences for ‘second’ assessments that affirm their prior
assessments (or negate them, in the case of self-deprecations). She describes
what she calls a ‘preferred-action turn shape’: marked by immediate response,
lack of explanation, delay or requests for repetitions of clarifications.

By contrast, she demonstrates ‘dispreferred-action turn shapes’ being
consistently marked by conversational phenomena such as pauses, explanations,
laughter and seeming agreements: ‘Yes but… no but…’, (to borrow from Vicky Pollard), softening a contradiction (or the ‘dispreferred’ affirmation of a self-deprecation).

Could Kant’s idea of ‘blame’, the expectation or demand that others should
agree with us be equated in some way to the idea of a ‘preferred’ and a
‘dispreferred’ assessment? The problem with mapping an essentialist idea onto a phenomenological framework like CA is that everything starts to look like either a chicken or an egg. Are the conversational phenomena products of some underlying rule, or are Kant’s observations of cases in which we ‘blame’ each other for disagreement and argue the point based on observation of certain degrees of dispreference in talk?

This may point to a fundamental problem in how I am constructing my question: philosophical discourses such as Kant’s and Hume’s may simply be incompatible with analyses based on the phenomena of dialogue. I suspect I need to read a lot more philosophy and conversation analysis to sharpen my questions up to the point that these very different kinds of source materials can be brought into play in a useful way.

XMLHttpRequest cannot load Origin http://mydomain.net is not allowed by Access-Control-Allow-Origin

Courtesy of Toby’s code for his BBC Stories visualisation for a demo we’re doing of our joint work at the DTC all hands conference, I had a day of cross-domain Ajax woe.

It was particularly annoying to run into this issue because I wasn’t even really trying to do cross-site AJAX, I just wanted to call some data from a SPARQL server running on a high port of my own server! But no, a different port, as far as the browser is concerned, is a different server.

After spending hours trying to “do it properly” and get Cross Origin Resource Sharing to work on my ISPconfig 2 debian lenny server, I just gave up.

I got it in principle, and I discovered that by adding Apache Directives like this:


Header add Access-Control-Allow-Origin "http://myserver.net"
Header add Access-Control-Allow-Origin "http://myserver.net:8080"
Header set Access-Control-Allow-Headers "X-Requested-With"
Header set Access-Control-Max-Age "60"
Header set Access-Control-Allow-Credentials true
Header set Access-Control-Allow-Headers "Content-Type, *"


To ISPConfig’s site control panel (instead of directly to Apache VirtualHosts), I did manage to get my headers doing the right thing:


saul@ni ~ $ curl -i -X OPTIONS http://mydomain.net/mydemo/

HTTP/1.1 200 OK
Date: Tue, 01 Nov 2011 14:38:56 GMT
Server: Apache (Debian) modpython Python modruby Ruby mod_ssl OpenSSL
Allow: GET,HEAD,POST,OPTIONS,TRACE
Vary: Accept-Encoding
Access-Control-Allow-Origin: http://mydomain.net
Access-Control-Allow-Origin: http://mydomain.net:8080
Access-Control-Allow-Headers: Content-Type, *
Access-Control-Max-Age: 60
Access-Control-Allow-Credentials: true
Content-Length: 0
Content-Type: text/html


at least as described in the various how-tos I was reading.

But after plenty of attempts, I just couldn’t get it working. Maybe it was something on the client-side that I just didn’t get. I’m no Javascript person…

Anyway, after battling hard to do it the right way, I caved and did it the sysadminny way, following the advice from Steve Harris I found on the 4store-support site in the first place and just set up a proxy to port 8080 so that the script could just request /whatever/ and get http://mydomain.net:8080/whatever/.

Bah.