---
# Associative Meaning in Language
If a symbol, a fashion item, etc. can have associative meaning, then why not language (a word, an expression, a pronunciation, a dialect)?
For instance,
ouch or
oops can be analyzed as a sound with some associative meaning.
--
Crucially, they are outside the compositional system and do not require an elaborate compositional semantic theory like \(\mathcal{L}_{\text{CI}}\) or its descendants.
Or even speaking a particular dialect might have associative meaning about e.g. social status, sexual orientation, etc. This is outside the compositional system too.
---
# Slurs and Associative Meaning
Similarly, ethinic slurs can be analyzed as involving associative meaning.
They have normal compositionally relevant meaning, as well as associative meaning.
E.g.
Jap has the same compositional meaning as
Japanese person but using this word attributes a certain attitude towards its user, just like certain symbols and hair styles.
Importantly, this associative meaning is not part of the compositional semantic system, and does not need \(\mathcal{L}_{\text{CI}}\), etc.
---
# Short Remarks on the Literature
Similar (if not identical) views in the literature:
- Anderson & Lepore (2013): Slurs are taboos, and inferences are brought about by violating the prohibitions
- Blakemore (2015): the offensive meaning of a slur "is derived from the meta-linguistic knowledge that the word is an offensive means of predicating and referring”
Recently formal linguists (e.g. Heather Burnett, Robert Henderson, Elin McCready) have been investigating how and why such associative meaning arises.
---
# Summary So Far
**Associative meaning** is a non-compositional kind of meaning, found everywhere, even outside language.
--
What Gutzmann calls Isolated Use-Conditional Items can simply be analyzed as involving associative meaning, and do not require a compositional analysis.
✅ |
isolated expletive |
ouch, oops |
✅ |
isolated mixed |
ethnic slurs |
❓ |
functional expletive |
damn |
❓ |
functional shunting |
yokumo in Japanese |
❓ |
functional mixed |
Japanese honorific verbs |
---
class: middle, center
---
# Functional (U)CIs?
For Gutzmann (2015), functional Use-Conditional Items are those that give rise to Use-Conditional Meaning by interacting with other words in the sentence.
--
E.g. Japanese honorific verbs
irassyat-ta
come.HON-PAST
-
At-issue: The teacher came
- **CI**: The speaker respects the teacher
If honorific verbs do have CIs, indeed we'll need a compositional system. But do they really?
---
# Honorifics: Pragmatic Flavor
Since Potts & Kawahara (2003) and Potts (2005, 2007) honorifics in languages like Japanese have been considered cases of CIs.
But are they? In particular, why are they not
presuppositions?
--
In fact, a use of an honorific sounds like the speaker takes it for granted that they respect the referent.
irassyat-ta
come.HON-PAST
'The teacher came (and I respect the teacher)'
(And honorifics pass the HWMT; data omitted.)
---
# Honorifics: Repeatability
Honorifics also do not show some of the properties that Potts (2007) claims are characteristics of expressives.
The honorific meaning is certainly
independent from the main utterance and
perspective dependent, and maybe
descriptively ineffable.
--
But repeating them do not necessarily strengthen the emotive content. Also, it's necessary to repeat them, whenever possible.
In this sense they are similar to gender features on pronouns.
---
# Honorifics: Displaceability
Honorifics also do not always take the maximum scope.
(15)
kono hito-ga
this person-NOM
gakusee-na-ra
student-is-COND
kyoozyu-na-ra
professor-is-COND
'If this person is a student, they might not come, but if they are a professor, they will definitely come.'
Here, the honorific attitude is conditionalized.
Gender features work in a very similar way here
(see Yanovich 2010).
---
# Honorifics: Quantification
Potts & Kawahara (2003) also claim that honorific verbs do not allow quantified subjects, and that's because of
Indepednence.
But their examples contain some confounds. A simple sentence like (16) is perfectly acceptable.
irassyat-ta
come.HON-PAST
'Most of the teachers came (and I respect them)'
--
Conclusion: Honorifics show very similar semantic behavior as gender features on pronouns and verbs, which have been analyzed as presuppositions.
---
#
Yokumo
McCready (2010) analyzes the Japanese word
yokumo as an expression that operates on the rest of the sentence and produces a CI without at-issue meaning.
'How dare you come here again!'
It seems that this is the only item that has been analyzed as a
shunting CI, i.e. it creates a sentence without at-issue meaning.
---
# Is
Yokumo a Pottsian Expressive?
Yokumo is unembeddable (as per most other performatives), so we cannot test (non-)displaceability.
We cannot test
Independence, because there doesn't seem to be at-issue meaning.
Perspective dependence holds, and
descriptive ineffability and
repeatability might too.
But notice that these three properties are true for normal expressions like
abominable and
very too.
---
#
Yokumo as a Speech Act Marker
Do we need to say that
yokumo is a CI?
I'm not sure. We know that natural languages have various Speech Act markers, for which linguists have developed non-CI based accounts.
- Exclamative constructions
- Tag questions, and various question particles, e.g.
desho in Japanese
- Assertive particles, e.g.
yo and
ne in Japanese
Why don't we say that
yokumo is one of them? They certainly need an analysis, perhaps in a compositional theory, but they don't seem to require the idea of CI.
---
#
Damn and
Fucking
We claim that items like
damn and
fucking also do not require a special compositional theory.
We analyze them as having associative meaning expressing the speaker's heightened emotion (anger, frustration, etc.) towards what they are saying.
(11)
John didn't ask if the damn dog was expensive.
Potts analyzed this as involving a CI that the speaker hates the dog.
According to our analysis,
damn here is just expressing the speaker's negative emotion towards what they are saying.
---
# What is the Negative Emotaion About?
The negative emotion epressed by
damn and
fucking often targets the referent of the phrase it modifies, but this does not need to be the case (Potts 2005).
(18)
I have to mow the damn lawn. (Potts 2005:7)
The speaker doesn't need to hate the lawn. Maybe she hates mowing the lawn, or having an obligation to mow the lawn.
---
class: middle, center
---
#
Bastard
For
bastard, the target of the emotion seems to be fixed to the referent of the phrase it modifies.
(19)
That bastard Connor was promoted. (Potts 2005:157)
We analyze
bastard as having both a presupposition and associative meaning.
- Presupposition: The referent is bad/vile/not nice
- Associative meaning: The speaker is angry/frustrated at what the sentence says
---
#
Bastard: Displacement
Bastard fails
Non-displacement in some examples:
(20)
I consider John a saint. But if he ever screws me over, I’ll crush the bastard like a bug. (Lasersohn 2007:12)
(21)
Whenever I pour wine, the damn bottle drips. (Potts 2007:171)
Our analysis can account for this (unlike the CI-based approaches).
- Presuppositions generally can take scope in modal contexts.
- Associative meaning is there (these sentences obligatorily involve heightened emotations)
---
class: middle, center
# Concluding Remarks
---
# Summary and Conclusions
Chris Potts's re-introduction of Conventional Implicature and his compositional theory \(\mathcal{L}_{\text{CI}}\) certainly had positive effects on the semantic community.
However, we think too many things have been analyzed in terms of CI, without a clear definition of what it is.
We are trying to develop a conservative view.
- **Associative meaning**, which is outside the compositional system, needs to be recognized.
- Re-analysis of CIs in terms of
at-issue meaning +
presupposition +
implicature + **associative meaning** without complicating the compositional theory.
---
# Honorifics and Gender Features
We claimed that honofirics should be analyzed in terms of presuppositions, just like gender features. They show similar semantic behavior.
But they tend to project more robustly than other types of presuppositions (and hence they have been called
indexical presuppositions since Cooper 1983).
We are trying to explain this by analyzing them as carrying presuppositions as well as associative meaning, just like
bastard.
---
# Selected References
- Anderson & Lepore (2013) Slurring words. Noûs, 47.
- Potts (2005) The Logic of Conventional Implicatures. OUP.
- Potts (2007) The expressive dimension. Theoretical Linguistics, 33.
- McCready (2010) Varieties of Conventional Implicature. Semantics & Pragmatics, 3.
- Gutzmann (2015) Use-Conditional Meaning. OUP.
- Leech (1981) Semantics: The Study of Meaning. Penguin.