How do you think we mean?

For the discussion of language mechanics, grammar, vocabulary, trends, and other such linguistic topics, in english and other languages.

Moderators: gmalivuk, Moderators General, Prelates

User avatar
Carlington
Posts: 1588
Joined: Sun Mar 22, 2009 8:46 am UTC
Location: Sydney, Australia.

How do you think we mean?

Postby Carlington » Sun Jul 05, 2015 1:15 pm UTC

There are a lot of different theories of meaning, and most of them are, almost by definition, untestable. Semantics is frustrating that way, in that it seems like a complete model of meaning cannot be obtained before a complete model of consciousness itself. I mean, we seem to think in meaning (at least, I know I do), so the two concepts, conscious thought and meaning, seem inextricably linked to one another. Nonetheless, there have been rather a lot of ideas put forth about how we might mean, and what it might mean to mean something. I'm not an expert - I've taken an intro level semantics course and done a little bit of reading around - but I've found that there are some ideas about meaning that seem nice and others that seem less so. I wanted a place to talk about that a little bit, and I'm quite interested to hear some other people's ideas and the reasons behind them.

Personally, I think that the best candidate for a proper theory of meaning is something akin to the theory of prototypes. That is to say, I think there's some sort of simulation of an ideal universe inside our minds, with which we conceptualise the ideal meanings of the language we perceive and generate. I think that the process of acquiring language is the process of assigning words to concepts, and in doing so we must necessarily distill out the most essential qualities of those concepts, the things that distinguish them from other concepts. This naturally looks like the process of creating boundaries between one concept and another, lines which determine whether something in the usually non-ideal real world should be categorised one way or another. So now we have some idealised model universe in the mind, populated by ideal representations of the concepts that we encounter, which we use to approximate the real universe full of real concepts. Of course, I don't want to claim that these concepts are static in the mind - I think that as we learn new meanings for words, and learn the subtleties and connotations of the meanings of words, the boundaries between concepts shift and change and the ideal representations, the prototypes themselves change too.

I recognise that this is a bit of a just-so story, and that there's really no justification behind it. I also find it odd that I should be attracted to an idea like this one, but be completely turned off by something like the idea of a natural semantic metalanguage (I dislike that theory enough that my final paper for the class was a categorical attack on it), when really the NSM theory is all about reducing things to their primitive representations as well. However, I don't think it's possible to generate a metalanguage capable of encapsulating all meaning in language - I think that doing so would require another metalanguage to establish the metalanguage, and then it's turtles all the way down. I'm also equal parts intrigued and disquieted by the successes that have been had in the field of statistical semantics, because I really don't like the idea that when I think I mean something, in fact my use of language is determined or described completely by looking at the language used immediately before and after. I don't like the notion that I'm just saying the things I'm saying because those are the things I'm statistically most likely to say. But even though I don't like it, the fact remains that it's brought a decent bit of success in terms of natural language processing.

But I'm really interested to hear what everyone else thinks. This seems to be the kind of thing where just kicking ideas around is the best way to make progress, and since I'm not nearly within reach of the big leagues of kicking ideas around (namely, publishing papers that people then cite and respond to), this seems like a very nice informal substitute.
Kewangji: Posdy zwei tosdy osdy oady. Bork bork bork, hoppity syphilis bork.

Eebster the Great: What specifically is moving faster than light in these examples?
doogly: Hands waving furiously.

Please use he/him/his pronouns when referring to me.

User avatar
eSOANEM
:D
Posts: 3649
Joined: Sun Apr 12, 2009 9:39 pm UTC
Location: Grantabrycge

Re: How do you think we mean?

Postby eSOANEM » Sun Jul 05, 2015 3:04 pm UTC

Prototypes are definitely a poor theory and it's fairly easy to show that it's an incomplete description of how people assign words to objects/how meaning works (in particular things like "tomato is a fruit, but it's not a fruit-fruit". Semantic fields are a much better field which extends the idea of prototypes.

Instead of simply defining terms as "things like [prototype]", we define a region of some semantic space to be described by that term. These edges are fuzzy (allowing for things that are "technically [term], but not a [term]-[term]" and the way different languages can have terms with the same prototypes, but describing different sets of objects. There's a big section about this (and why semantic fields are better than prototypes) in one of the language construction kit books by mark rosenfelder.

Edit: I'll see if I can dig out the section in a bit (probably tomorrow or the day after)
my pronouns are they

Magnanimous wrote:(fuck the macrons)

User avatar
Carlington
Posts: 1588
Joined: Sun Mar 22, 2009 8:46 am UTC
Location: Sydney, Australia.

Re: How do you think we mean?

Postby Carlington » Mon Jul 06, 2015 10:22 am UTC

Funnily enough, it was a post you made that got me thinking about this again, so it's good to have you in this thread!

I didn't actually realise that there were any semantic theories anything like semantic field theory. As it turns out, the way I conceptualised meaning all through that intro course was pretty close to how the idea of semantic fields seems to have started out. Eventually, I gave up on the idea as silly, because we didn't really go into much depth and I don't remember us touching on semantic fields at all, so I assumed that it mustn't exist and that I must be wrong (if I come up with an idea and it's not talked about, I usually assume there's some gaping flaw that makes me obviously wrong, and that I'll figure it out eventually - that usually ends up being the right course of action, anyway, and it's also been borne out by the number of times I've asked questions in a class only to have some gaping flaw pointed out to me.) So eventually, I kind of reduced the idea back down, and figured that it was basically an extension of prototypes, and that what I was thinking of was probably really not that different from prototypes at all - which is why I talked about them here. I think there was a little bit of holdover from what I now know is semantic field theory, in where I talked about creating boundaries between one thing and another, and the boundaries shifting and changing over time.

It'd be really cool if you did dig that section out, but I sent an email to my professor from last year as well and got some recommendations on good places to start reading, so even if you don't I'll have something to look into, which is a nice feeling.
Kewangji: Posdy zwei tosdy osdy oady. Bork bork bork, hoppity syphilis bork.

Eebster the Great: What specifically is moving faster than light in these examples?
doogly: Hands waving furiously.

Please use he/him/his pronouns when referring to me.

Aiwendil
Posts: 311
Joined: Thu Apr 07, 2011 8:53 pm UTC
Contact:

Re: How do you think we mean?

Postby Aiwendil » Mon Jul 06, 2015 2:45 pm UTC

Not enough time to say anything substantive here at the moment, but I do want to make one point. I think that there are several related but different questions that sometimes get conflated under the heading of 'theory of meaning'. In particular, I think a distinction must be drawn between questions like 'how does a proposition mean something about the world?' and questions like 'how do people cognitively interpret signs and locutions?'

User avatar
eSOANEM
:D
Posts: 3649
Joined: Sun Apr 12, 2009 9:39 pm UTC
Location: Grantabrycge

Re: How do you think we mean?

Postby eSOANEM » Fri Jul 10, 2015 1:14 am UTC

The language construction kit by Mark Rosenfelder wrote:
The structure of the lexicon
A key insight of Ferdinand de Saussure was that the problem of single-word meanings could be sidestepped by looking instead at language as a structure. Meanings don't exist in isolation; they're circumscribed by their relationship with other words.

{some explanation of the ways words can relate to others}


Reference mismatch and new senses
Somewhat more ambiguously, speakers might use a word differently from its meaning, till the meaning changes. In terms of our metaphor, we could say the meaning is remapped to fit its usual referents.

{diagram}

...

Some oddities of meaning
...

Another naive formulation is that all the members of a class have "something in common", even if we are not sure what it is; Ludwig Wittgenstein pointed out that this is not the case for many words, such as game, whose instances have family resemblances but do not all share a set of features.

...

Categories and prototypes
Not all members of a class are equal. Eleanor Rosch and George Lakoff have emphasized that classes have prototypes or central members; speakers consider these typical examples of the class, list them first if asked for examples of the class, and think about the class in terms of prototypes. The prototype bird, for isntance, is probably a songbird- small, flying and untamed. General statements about a class may be true only of the prototype (e.g. birds fly)


In the advanced language construction kit (effectively an extension including a bunch of stuff he left out of this one) he also says that prototypes can often make thinking about things easier. It seems pretty clear that, whilst they're useful, they're far from the whole story (and thinking about the distribution of regions in a semantic field is a better metaphor).
my pronouns are they

Magnanimous wrote:(fuck the macrons)

User avatar
quintopia
Posts: 2906
Joined: Fri Nov 17, 2006 2:53 am UTC
Location: atlanta, ga

Re: How do you think we mean?

Postby quintopia » Fri Jul 10, 2015 2:50 am UTC

Personally, I am in the Rosch/Lakoff camp. Women Fire and Dangerous Things made a hell of a lot of sense to me. And the experiments involving basic-level categories must be pointing to something. Ergo, ideas metaphorically linked to each other and to a small number of semantic prototypes has to be the way to go. The theory actually already accounts for eSOANEM's complaint. Some things have more things in common with the prototypes than others, some members of a category are more central than others, and we can measure the "prototypicness" by counting the number of metaphorical inferential steps or properties shared with the main prototypes.

One issue I'm surprised that eSOANEM does not raise is how context modifies categories and prototypes, for instance how in a botany context, no one is going to say "rose hips are technically fruits", because they are canonical category members in that context. They just are fruits. In a food context, however, people are more like to say "eh, I guess they're fruit? you can make jam with them right?" My best guess as to how this fits in is that there never were strict boundaries between mental categories, and the prototype you keep in mind shifts with the context, so that, in one context, a thing that is normally gestalt-wise metaphorically far from the default (basic level) prototype ends up being much closer to the prototype adopted for that context, and therefore seems a legitimate member whereas before it was a stretch.

User avatar
Carlington
Posts: 1588
Joined: Sun Mar 22, 2009 8:46 am UTC
Location: Sydney, Australia.

Re: How do you think we mean?

Postby Carlington » Fri Jul 10, 2015 7:48 am UTC

More things to read! Excellent! That's beginning to sound closer to things that make sense for me as well. Regarding your rosehip example, would we be willing to go as far as to say that the difference is enough to constitute two different senses of 'rosehip'? It seems like maybe it's a little more subtle than that, but I also think it's quite heavy-handed to say that entire regions of wordspace are shifted to and fro as context changes.
Although, with that said, when we consciously make an effort to communicate in a context-specific way with context appropriate language use, maybe we do temporarily redefine the boundaries and distances between words.

Jargon is almost like another dialect - certainly it has a much smaller measure of mutual intelligibility than would be found between the language used between two ordinary speakers in a language community. Do you think that maybe technical senses and differing definitions are a kind of jargon? In a similar vein, what happens to the internal representation of meaning when a bilingual speaker switches language? Certainly some words/regions ought to stay the same, and just be relabelled (the shape of the definition for 'apple' probably looks very much like the one for 'pomme' and for 'Apfel') but some new regions will undoubtedly be formed, and some will vanish (there's probably overlap between 'goluboy' and 'blue', or 'sadness' and 'saudade', but they certainly don't fill the same region).
Kewangji: Posdy zwei tosdy osdy oady. Bork bork bork, hoppity syphilis bork.

Eebster the Great: What specifically is moving faster than light in these examples?
doogly: Hands waving furiously.

Please use he/him/his pronouns when referring to me.

User avatar
quintopia
Posts: 2906
Joined: Fri Nov 17, 2006 2:53 am UTC
Location: atlanta, ga

Re: How do you think we mean?

Postby quintopia » Fri Jul 10, 2015 8:19 am UTC

I would suggest that the metaphorical category space is fixed, and it is only the portion of a category that we are paying attention to and primed for that changes with the context. I think this is what happens as soon as we decide to communicate in a context-specific way.

I agree with your assessment that jargon is a dialect. Jargon typically just means "using a more exclusionary synonym in a context where it is understood that your audience will understand the uncommon phrasing". I think translating things into jargon is largely more conscious (because it requires more thought given to the context) than speaking another language. Someone fluent in a language doesn't have to think as hard to speak as, say, an actor speaking a role outside his natural dialect.

I reckon you are right that category space largely remains constant when an individual learns more than one language, but I also think that averaging over individual mental category spaces in a population will definitely yield cross-cultural differences. Another example of adaptation to context writ large. In particular, I suspect that no matter what language you are speaking, you are drawing from a slightly modified category map for your native language. But this is all wild speculation, as I've never seen empirical support for this exact proposition. (Though there is plenty of support for the above claim of cross-cultural differences.)

User avatar
eSOANEM
:D
Posts: 3649
Joined: Sun Apr 12, 2009 9:39 pm UTC
Location: Grantabrycge

Re: How do you think we mean?

Postby eSOANEM » Fri Jul 10, 2015 10:55 am UTC

quintopia wrote:Personally, I am in the Rosch/Lakoff camp. Women Fire and Dangerous Things made a hell of a lot of sense to me. And the experiments involving basic-level categories must be pointing to something. Ergo, ideas metaphorically linked to each other and to a small number of semantic prototypes has to be the way to go. The theory actually already accounts for eSOANEM's complaint. Some things have more things in common with the prototypes than others, some members of a category are more central than others, and we can measure the "prototypicness" by counting the number of metaphorical inferential steps or properties shared with the main prototypes.

One issue I'm surprised that eSOANEM does not raise is how context modifies categories and prototypes, for instance how in a botany context, no one is going to say "rose hips are technically fruits", because they are canonical category members in that context. They just are fruits. In a food context, however, people are more like to say "eh, I guess they're fruit? you can make jam with them right?" My best guess as to how this fits in is that there never were strict boundaries between mental categories, and the prototype you keep in mind shifts with the context, so that, in one context, a thing that is normally gestalt-wise metaphorically far from the default (basic level) prototype ends up being much closer to the prototype adopted for that context, and therefore seems a legitimate member whereas before it was a stretch.


Pure prototype theory doesn't account for languages which have the same prototype but include different things. Clearly the prototype isn't the whole sotry.

The context thing isn't particularly important to me, you've effectively got several different definitions (region boundaries) of the same word and context tells you which one to use. Ultimately, that doesn't change what the meaning is.

Additionally, the statement that "birds fly"can be considered a true statement but pretty much anyone would agree that an ostrich is a bird also makes it clear that we can't get the whole truth by just considering prototypes.
my pronouns are they

Magnanimous wrote:(fuck the macrons)

User avatar
quintopia
Posts: 2906
Joined: Fri Nov 17, 2006 2:53 am UTC
Location: atlanta, ga

Re: How do you think we mean?

Postby quintopia » Fri Jul 10, 2015 3:18 pm UTC

I'm not sure what you mean by pure prototype theory, but the fact that different cultures think in different metaphors would seem to me to fully account for the same-prototype-different-category-boundaries problem.

If I ask someone "Do birds fly?" they will surely interpret it to mean either "is flying a property associated with the prototype idea cluster 'bird'?" Since the prototypical bird does fly the answer will be a quick yes. The answer would be yes also if the question were interpreted to mean "are there birds that fly?" The only situation where the answer could conceivably be no is if one were to assume a universal qualifier is involved, but since you did not ask "Do all birds fly?" why would anyone assume that is what you meant?

User avatar
eSOANEM
:D
Posts: 3649
Joined: Sun Apr 12, 2009 9:39 pm UTC
Location: Grantabrycge

Re: How do you think we mean?

Postby eSOANEM » Sat Jul 11, 2015 12:12 am UTC

I mean a model in which the prototype defines the term. Different fundamental metaphors can certainly get you some way but I'm not sure it's sufficient.

The birds question raises two points:

1. People think about meaning in terms of prototypes.
2. That this is an incomplete description. "Yes" is not a completely accurate answer; "some do" is the answer which provides completely accurate information but is only the natural response if people conceptualise meaning as a region in some semantic space (the fact this is not the most common response, makes it clear that this conceptualisation is uncommon).

How people believe meaning works and how it does need not be the same however. The fact people who said "yes" to the question "do birds fly" will also say "yes" to the question "is an ostrich a bird" without equivocation seems to me to be good evidence that there's more going on than simple comparison to prototype.
my pronouns are they

Magnanimous wrote:(fuck the macrons)

User avatar
quintopia
Posts: 2906
Joined: Fri Nov 17, 2006 2:53 am UTC
Location: atlanta, ga

Re: How do you think we mean?

Postby quintopia » Sat Jul 11, 2015 2:54 am UTC

In both of the questions "do birds fly?" and "is an ostrich a bird?" there is no thought involved. In both cases you answer by regurgitating memorized facts. (If you don't believe this, show a two-year-old who has never encountered one a penguin/ostrich/duck and ask them whether it is a bird. I'd say odds are better than even you get "no".)


Return to “Language/Linguistics”

Who is online

Users browsing this forum: No registered users and 7 guests