erratio | Entries tagged with cogsci

In case you still had the misconception that you're able to perceive reality as it really is, try taking a phonetics course. Specifically, try to learn how to transcribe speech.

As a student of linguistics, I of course already knew that the brain does a ton of work converting sound streams into recognisable speech. I already knew that speech sounds are affected by their surrounding environment, and that my brain does cool compensatory stuff and predictive work to let me easily pick out individual words and meanings with remarkably few mistakes.

But a couple of days ago we did our first real-time transcription exercise in class where someone reads out a bunch of words in isolation, and those things are *hard*. That sound at the end of the word, was it an unreleased [t]? A glottal stop? Nothing at all? Something else entirely? That word pronounced in an American accent that sounds kind of like 'cut' to me, was actually 'cot', and if there'd been any context I would have easily gotten the right word and wouldn't have even noticed that the vowel sounds more like [u] than [o] to me.

Basically what I'm saying is that I now have a new visceral appreciation for how hard this stuff is.

Other classes I'm taking this semester include:
How to use Google and Twitter searches as totally legitimate sources of linguistic research, (aka Experimental Syntax)
Assigning stress to words and phrases is way harder than you would naively expect (aka Foot Structure)
Fancy mathematical formalisms you can use as a framework to understand syntax instead of making up plausible sounding crap that doesn't really engage with most of the previous literature (aka Compositional Syntax)

Crossposts: http://erratio.livejournal.com/137724.html

ok, so this isn't actually liveblogging, I read the first third of the book a couple of weeks ago. Anyway, on to the interesting bits, in no particular order.

* Supposedly, laughter/amusement is an exaptation of the flight response (panting to get more oxygen), frisson (enjoyable chills) comes from the fight response (erecting all your hairs, cat-style), and a gasp/sense of awe comes from the freeze response (one large breath to get as much oxygen as possible, followed by holding your breath so as not to attract attention).

* The book started out as being about the psychology of music and ended up being mostly about the psychology of expectation, albeit with mostly musical examples.

* we have models of how music ought to go based on our previous music knowledge/experiences, but also based on on-the-fly modelling and possibly a general idea of melody and closure and so forth - a betting game where people had to predict the next note for a traditional tune of some non-Western variety showed the group who were familiar with that music did way better than a group of Western music students, but the students still as a group still did well above chance

* There are lots of ways to try to test people's expectations of music. None of them are even close to ideal. They include ERP studies, asking people to improvise the next note, getting people to predict or bet on next possible notes, playing them a bunch of probe tones and getting them to pick the best, head-turn studies in babies...

* the main findings of these studies have been that most people do have expectations of how music ought to go, including that the melody line should on the whole descend during the second half of the piece, a large jump up or down the scale should be followed by movement in the opposite direction, for intervals between adjacent notes to be smaller rather than larger, and for small intervals to generally all move in the same direction. Most of these were tested cross-culturally too. However, these expectations don't actually hold for music (again, cross-culturally). Rather than a large interval being followed by movement in the opposite direction, we actually see regression to the mean. Rather than small intervals generally all being in the same direction, the tendency is for small intervals to move down the scale. And musical phrases follow an arch, not only descending in the second half. So this shows that listeners are using heuristics to form their expectations rather than building totally accurate models.

Crossposts: http://erratio.livejournal.com/136791.html

ok, semester is over and obviously the humour series sort of got away from me. So instead of trying to continue it in detail, I'll offer a summary of the key points

The theory offered in Inside Jokes is this: we are constantly constructing new mental models of our environment (mental and physical) on the fly, based on knowledge and guesses. It is important that these models be as accurate as possible - that tiger hidden in the bushes better be recognised as a tiger or you won't survive long. Amusement is your brain's method of getting you to troubleshoot your models by acting as a reward mechanism, kicking in when you discover that a part of your model that you'd accepted as accurate turns out to be wrong. The urge to laugh is the reward/we know we are amused because we want to laugh. Jokes and deliberate humour are a super-stimulus for the amusement system.

To that I would add: the reason amusement is signalled by laughter, as opposed to some other mechanism, is that laughter was co-opted from an earlier fear response. Huron (from Sweet Anticipation, which I'm reading now) suggests that we can trace laughter from the play-panting common to primates and some mammals, which in turn used to just be hyperventilation in preparation for a flight response. The panting/hyperventilation came to be a signal of low status in response to being confronted with a higher status peer (you can tell I'm afraid of you because I'm panting), which then got co-opted for play, and in humans became more obvious (vocalisation) and efficient (only using the out-breath) in the form of laughter.

The previous paragraph is an ev-psych just so story, but I like the way it ties in all the status/signalling aspects of laughter. It suggests that laughter has two distinct purposes: displaying status, and troubleshooting mental models. It explains why we feel as if higher-status people come across as funnier - we laugh as a signal of low status and then misinterpret the laughter as a sign of being amused, particularly since the environment suggests that being amused is a plausible response (unlike the guy who suffered from involuntary laughter, who did not feel amused). This also explains why it's so hard to account for all kinds of humour/laughter using a single theory.

Not adequately covered by either half of this model: laughing at other people, particularly people who are lower status than you. Although it might be as simple as 'my model which previously suggested that this person was intelligent/competent was incorrect'. In-jokes, which I'm going to explain as a form of self-anchoring - I found this funny in this context in the past and so now I'm the kind of person who finds this joke funny when said in a similar enough context.

Next: Sweet Anticipation, by David Huron. It's a book about the cognitive science of music and also of expectation. Instead of reading the whole thing and then attempting to summarise it I might try something more like live-blogging, where I stop every now and then to summarise the interesting points so far.

Crossposts: http://erratio.livejournal.com/136667.html

Some background on Nicaraguan Sign Language (NSL):
NSL is about 30 years old now. It began when a special school for the education of deaf children was established in Managua, allowing deaf children starting from age 4, who'd previously only had very ad hoc systems of home sign with their family, interact with 25-30 other deaf kids, each of whom brought their own idiosyncratic home sign system/language with them. When those kids interacted they created a completely new, richer sign language. And then since it's a school, every year another cohort of 25 or so new kids would come in, and those kids in turn expanded on the original system created by the first few cohorts, and so on until NSL reached its current status where it's basically a full-fledged language, created within the last 30 years from virtually nothing and without major contamination from other languages. Basically, it's a linguist's dream language, because we have detailed records of what the language looked like at each stage of growth (more on this in a moment) and so we can literally see the grammar unfolding over time rather than having to guess, like we do for pretty much every other language. And it turns out that we can also see what cognitive functions language is and isn't necessary for, which is pretty cool. More on this in a moment too.
The way this school works is that kids go from age 4 until age 14, at which stage they graduate and henceforth are allowed to hang out at the deaf club, but are (obviously) no longer at school. There are also often older kids just starting who couldn't come previously, and their language development is obviously not as good since they haven't had access to a decent source of language until that point. The early cohorts didn't really hang out with each other outside of the school context, but the later ones, having grown up in the age of cell phones, do. At school, the kids get several hours a day to hang out with each other - during food/play times, on the school bus (some of them live as much as 2 hours away, so that's a lot of time to socialise with the other kids), behind the teacher's back... The schooling is pretty much all done in Spanish and mostly with non-signing teachers. As you might expect, not a whole lot of regular school learning actually goes on, although more recently they've started hiring adults from the first generation of signers as teachers so they can actually communicate. Plus, texting via cell phones means the kids are way more incentivised to learn to read than the first generations were.

On sign languages in general: when signs are coined they are often iconic in some way or other. For example, the sign for a king may be the action of putting a crown on, or the sign for a cat might be drawing imaginary whiskers on your face. But there's nnothing principled about what iconic aspect of a thing or action will become encoded as a sign, and signs tend to get less iconic over time.

So, Ann Senghas. She's been going down to this school for the deaf every summer for many years now, documenting their language, getting them to complete various linguistics tasks, and so on. And now, onto the pithy details of the talk, listed in bullet point form as usual because I'm lazy and can't be bothered with trivialities like "good writing".

* The NSL signers can be split into roughly 3 generations, descriptively called first second and third. First generation started school in the 70's, second in the 80's, third in the 90's
* If you look at a video of each generation signing, there aren't any obvious differences at first, except in speed - first generation is slow compared to second is slow compared to third. But they're all clearly using language, not pantomiming or gesturing.
* However if you look more closely, there are bigger differences. Two ways that we saw today included the expression of number and expression of space. Others that were mentioned include expression of path/manner of movement, syntax, theory of mind stuff, and general 'with it'-ness
* On path/manner of movement: where the first and second generation would express a ball rolling down a hill by more or less pantomiming an object rolling down, the third generation would express a ball rolling down a hill by first indicating a rolling thing and then indicating a descent.
* On syntax: for the earlier generations, verbs could only take a single argument each, so "the boy fed the woman" would be expressed as "woman sit; boy feed"
* On expression of number: the first generation would express number the same way us non-signers generally would: 15 would be 5 on both hands followed by 5 on one hand. The second generation developed a more efficient (one-handed, faster) system that builds on that of the first generation: A girl counting to 10 counting the first 5 normally on one hand followed by counting from 1-5 again on the same hand but accompanied by a slight twist. Another girl asked to express the number 15 did so by first indicating a 1 and then moving her hand slightly to one and then indicating a 5 (so basically a 1 in the 10's column and a 5 in the units). Kids in the third generation came up with a new system altogether that loses a lot of the transparency but is even faster and more compact: 15 is expressed by holding the middle finger towards the palm with the thumb (imagine you're trying to form a ring with your thumb and middle finger - this represents 10) and then flicking it outwards to show 5 fingers. Apparently the older generations understand these kinds of signs but are disdainful of them - "they don't even look like numbers, it's just a flick!". This kind of pattern exemplifies the different generations: first generation functional but not particularly efficient, second generation has some kind of systematic improvement that allows them to express themselves more efficiently, and third generation as often as not will come up with something way more abstract that bears very little iconic resemblance to its meaning.
* On expression of space: there's a task linguists sometimes get people to do that goes as follows: person A has to describe a simple picture to person B, who then picks the matching picture on their side of the test area. In this case the pictures were of a tree and a man, where the man would be standing either to the left or right of the tree and could be facing towards or away from the tree, or out to the audience or away from it. Ann Senghas gave this task to her signers to find out how they expressed spatial concepts. Instead what she found was that the first generation failed the task - they couldn't encode spatial relations and performed at chance. In the later generations everyone could do it just fine. We were shown a video of the task being done by a first generation speaker and her third generation nephew, where during a break in the task she asked him to explain how to get it right. The kid does a pretty good job of explaining something that must have seemed ridiculously obvious to him - if the person is on this side of the tree then you put them like so, otherwise you put them on the other like so. This isn't something you can practise, you just look and then do it. Easy! (very rough paraphrase from memory). She did not get it.
* On theory of mind and 'with it'-ness: the first generation fails at second-order theory of mind, aka situations where you have to express what you know that I know. They're also a lot less 'with it' in general - like when Senghas is trying to coordinate with them for meetings and such, they're just a lot less good at it. They're also way less good at metalinguistic stuff - being aware of how you express things.

Crossposts: http://erratio.livejournal.com/135089.html

There's been some interesting talks lately, but today was the first one in a while that made me think "I should blog about that". But since I also would like records of the other talks, I'm going to start trying to summarise the ones I found interesting.

Julie Van Dyke - on language processing using cue retrieval
* Language processing is really heavily dependent on working memory.

* But we don't actually know much about working memory (eg. how much of it we have), so to be safe let's assume that a hypothetical person can only remember the last item they heard/read. This isn't as insane as it sounds - computer models have indicated that processing can do pretty well even with such an impoverished working memory. Everything that isn't in active working memory is absorbed passively and can be called upon (albeit not as easily)

* So let's consider a few hypothetical sentences: 1) the book ripped 2) the book recommended by the editor ripped 3) the book from Michigan by Anne Rice that was recommended by the editor ripped. How does a listener tell if 'ripped' forms a grammatical sentence with 'the book'? There are a few ways: they could search forwards or backwards through the sentence, in which case you would expect processing times to reflect the amount of material between "the book" and "ripped". Or you could do cue-based retrieval, where you filter the sentence for words that have the features you're looking for, in which case you wouldn't expect there to be significant time difference in retrieval. As the name of the talk might suggest, people use cue-based retrieval.

* So now we have a model where we store words as bundles of semantic/phonological/etc features and then retrieve them by using those features. But what if the sentence has several possible words that have the features you're looking for? In that case, retrieval might get blocked due to interference from the other items. This, according to Julie Van Dyke, is why people forget. (I don't know whether she meant in general or when processing sentences. Hopefully the latter)

* And the main difference between people who are good at processing (eg. fast readers) vs those who aren't, is almost entirely based on how detailed your representations are. Because if your word representations are super detailed with lots of features, it's easier to zero in on them. And, good news, the main factor in how good your representations are (after controlling for IQ and a bunch of other bothersome details) is practice. So if you suck at reading, all you need to do to fix it is read more.

(preamble: anyone who's a regular reader of LW can safely skip this post, it's nothing that hasn't been covered there a few hundred times)

There's a folk psychology idea that emotions and wisdom are opposed traits. There are lots of people who make really short-sighted impulsive decisions based on their emotions who would obviously benefit from stopping once in a while to think through the consequences of their actions. And on the other end of the spectrum is the Spock stereotype that most nerds are haunted by at some point or another**. Well, good news everyone! Turns out there's no dichotomy between the two! In fact, you need both!

Let's pick on Spock for a moment, and take the kind of scenario he might be faced with in a typical episode of Star Trek: there's a couple of crew members down on a planet who've been captured by the local bad guys. Those crew members will die if they're not rescued. Only problem is that they're being held in the middle of the bad guys' Fortress of Doom, and according to Spock's calculations a typical rescue attempt only has a 5% chance of succeeding and has a 50% chance of resulting in the deaths of the entire rescue team. What's the rational thing to do here? *

What if one of the crew members being held is Scotty, who they need to keep the ship running? What if it's Captain Kirk, who they need to seduce alien queens**? Is it more rational to mount a rescue then? Why? It's not like any of the numbers of the original estimate have changed. Dig into Spock's 'rationality' and it pretty clearly comes down to number of lives saved. A rescue attempt with a 5% chance of success and a 50% chance of more deaths is a lousy gamble. The perceived odds shift (even though the bare numbers haven't changed) when taking into account more important crew members because those people are essential to preventing more deaths further down the line. But why is it rational to save lives?

The real answer here is that Spock isn't actually ignoring his emotions at all. The only reason anyone would be interested in saving lives is if they value life over death. To unpack that further, we like it when people are alive and we don't like it when people die. Or maybe you do like it when people die but don't like it when everyone shuns you because you're a creepy death-loving weirdo, so you pretend to dislike death. The point here is that ultimately you act according to your values, and your values consist of emotional valencies towards certain concepts, eg. +10 life, -10 death, -20 being alone forever, +5 having a prestigious career, and so on. Without values, you have no mechanism to decide that thing A is a better decision than things B-Z. Some of these values are more common and deep-rooted than others, mostly because we only really have a small number of things we like and dislike, and so a value like "having a prestigious career" (which can change when you re-evaluate your life) is just a fancier version of "being liked by others" (which is much harder to shift and can be satisfied in lots of different ways).

Transient emotions can also affect our values. Dan Ariely, in his book Predictably Irrational, talked about some experiments on how arousal affects decision making. A bunch of young men were asked questions like "would you have sex without using protection?" and "would you enjoy being spanked?" while in a normal baseline state. Unsurprisingly, they all said they would always use protection, wouldn't engage in taboo or kink, would always get consent, and so forth. Then they were given a stack of porn and given similar questions while they were aroused, and lo and behold, suddenly things like consent and protection were less important. Not because they were originally lying***, but because arousal causes a temporary rearrangement of your values to encourage you to procreate.

This is getting longish, and I have a roleplaying game to go to, so I'll stop here. Next post will be about curiosity, humour, and the evolutionary importance of having good mental models.

* I should probably mention that I've watched very little of the original series, and it's been a long time since I wached any of The Next Generation, so really I'm just making stuff up here.

** ok fine, and also to get into punch-ups. And I suppose to command the ship occasionally

*** even if their original answers were just signalling, I would argue that that's still a strong indication of their values: namely that their actual values around sex were getting outranked by their values around appearing virtuous, and then arousal changes that ranking****

**** One of my current classes is all about analysing phonology using a ranking system called OT. I feel a bit like I have rankings on the brain as a result

Crossposts: http://erratio.livejournal.com/133233.html

Last week we had a presentation by Keith Chen, who you may remember as the economist/management guy who caused a bit of a furor by claiming that the way any given language marks the future tense is a good predictor of future-oriented behaviours like saving money, taking care of health, and that sort of thing.

Some background: If you know about hyperbolic discounting it's probably a no-brainer to hear that if you give people a 401k/superannuation form with a picture of themselves now they'll put down lower payments than if you present them with a photo of themselves photoshopped to look old. The reason for this is that as a species we like to enjoy the good times now and make our future selves pay for it, but if you make people identify more with their future selves then they don't feel quite so great about putting stuff off that way.

With that out of the way, Chen's presentation can be summarised more or less as follows: Languages vary in the way they mark the future, with some languages like Hebrew forcing you to express the future explicitly whenever you're talking about future events and other languages like Chinese letting you talk about the future without using overt tense markers. For languages like German or English which have both overt (I will eat breakfast tomorrow) and covert (I am going to eat breakfast tomorrow) ways of talking about the future that can be used more or less interchangeably, the linguists who did these surveys looked at frequency of each type and and categorised the language as weak future-marking or whatever based on the most frequent forms.

The modern version of the Sapir-Whorf hypothesis says that language structure has weak priming effects on speakers. For example, in languages where the grammatical gender of "bridge" is feminine, people are more likely to describe them as "graceful" or "soaring" whereas speakers of languages that have masculine bridges are more likely to describe them as "strong" or "durable". (See Lena Boroditsky's publications page for lots more experiments of this type).

Chen used a bunch of census type data that included the language spoken at home to find that even after you control for a fairly impressive array of confounding factors, there was a strong correlation between savings behaviour and type of future marking in a language. In fact, he claimed, language type is a better predictor of savings behaviour than a bunch of other factors that economists usually consider to be pretty important, including trifling considerations like the country's economy.

Overall story: Every time you use an overt future marker you are priming yourself to think of future-you as a different person to present-you, so you're more likely to do what makes present-you happy. Or alternatively, there is a common cause for the savings behaviour and speakers' choice of how they talk about the future.

Crossposts: http://erratio.livejournal.com/132108.html

Benign violation (BV) theory

There isn't actually that much to say here that doesn't properly fit into either IR theory or status/superiority theory, but here goes.

The central premise of benign violation theory is that humour exists in the violation of social norms as long as there's no real harm in it. The main point in favour of it is that it describes accurately why things like slapstick and verbal sparring can be funny but attempted murder and arguments aren't, even though they mostly involve the same actions. It also captures why so much humour revolves around sex, excrement, and death, all of which are things you don't talk about in polite company.

Minsky (1981) proposed a sort of Freudian account of humour. Namely that your brain has a bunch of cognitive censors designed to taboo certain kinds of words/thoughts such as sex or excrement related, a la Freud, but also censors for faulty reasoning. And then cheating these censors is 'naughty' and this is what you find funny, successfully carrying out taboo acts or thoughts. This sort of provides an explanation for wordplay humour, since the joke usually lies in an ambiguity between a normal serious reading and the incorrect nonsensical one.

Another point in favour of BV theory is the evolutionary psychology explanation of laughter. Some types of primates have a 'false alarm' signal to go along with the 'snake', 'jaguar' and other assorted predator signals. And on top of that, apparently when chimps play they make a special 'play face' and engage in a kind of panting, which both help to signal that they're playing and the situation isn't serious. Hurley, Dennett, and Adams' explanation of laughter is as a sort of combination of these things: a way to signal that there's no real danger and I'm just playing with you. And so since humour is accompanied by the 'not serious' signal, the logic is that humour can pretty much be characterised as potentially-harmful things done in a non-harmful manner and that's why we all laugh at it.

A final point in favour of BV theory is that it accurately captures the intuition that it's difficult to be in a negative emotional state and find something funny at the same time. But if that negative state isn't due to the potential humour, BV theory says nothing about why I should find something less funny then than when I start off in a neutral or positive emotional state.

What BV theory can't capture

1. Humour isn't always harmless. See: pejorative and bullying humour, mean humour, satire, humour based on inferiority of others (eg. Irish jokes)

2. All the subtleties of humourcraft: if humour is just being non-serious or cheating an internal censor, it should be much easier to craft hilarious jokes than it is. All I would need to do is go out to a public space and say 'poo' a lot, or do something obviously nonsensical to trip the 'faulty reasoning' censor. Or for that matter just lie in bed and think of nonsensical or scatological scenarios. Jokes shouldn't really get more or less funny depending on whether you've been exposed to them before, since a norm violation isn't going to be less of one over time.

Overall, benign-violation theory makes a decent attempt to provide an explanation of humour but misses the mark on many many levels.

Crossposts: http://erratio.livejournal.com/131636.html

Incongruity resolution (IR) theory

The central idea of incongruity resolution theory is that when we have an expectation that is suddenly resolved, we find it funny. IR theory comes in many different flavours: Kant claimed humour is when we have 'strained expectations that come to nothing", other modern researchers have claimed that it's when we develop two competing frames/expectations from a setup, which is then resolved in favour of one by the punchline, or that we have one frame for the setup and another for the punchline and the humour comes from resolving the two, or that it's when our perceptions and our abstract representations clash, or any number of other variations that involve unexpectedness. And not just any old unexpectedness - pretty much everything that happens to us isn't anything we actively expect. The kind of unexpectedness IR theory calls for are things that we expected *not* to happen as opposed to things that we merely weren't expecting. I didn't expect to see the particular guy at the library who checked my books out for me today, but if he'd been dressed up as Death I probably would have found it amusing.

What incongruity resolution theory gets right:

It accounts for why watching people fall down is widely considered hilarious. It explains most wordplay (where the incongruity comes from ambiguity in meaning). It somewhat explains parodies and obscure humour, where the requirement of being able to draw on your previous knowledge is likely to bring a set of expectations with it to be shattered. It explains unhelpful humour, since we have expectations of people saying things to us that are relevant and truthful (see Grice's maxims for more detail), and to a certain extent mean humour (by the same maxims I expect people not to be unnecessarily mean).

What incongruity resolution theory gets wrong:

There are lots of examples of incongruousness that aren't funny. Some examples: a patient with baffling symptoms, lies, mysteries and puzzles, snow out of season, an instrument out of tune at a concert.

There are plenty of jokes that remain funny even when you already know the punchline, including ingroup humour and really good comedy movies and shows like Monty Python or the earlier seasons of the Simpsons. In these cases, there aren't any expectations being proven false or resolved in an unexpected way, since I already know what's going to happen

IR theory also does a bad job of explaining the social aspects of humour - why other people's laughter makes things funnier, although we could stretch the theory to cover it by guessing that other people laughing makes you more likely to reach the same interpretation as them and therefore also find it funny. Finally, there's still a lot of vagueness in the theory: what is incongruity exactly? Most of the proposed definitions contradict each other. IR theory is more of a description than an explanation.

Crossposts: http://erratio.livejournal.com/131360.html

It occurs to me that I may have misrepresented part of the status/signalling/superiority theory of humour, in that I focussed on why you would tell jokes but only very briefly mentioned why it is that we might find things funny even when they have no obvious author. So just to make things clear: another way to state the status/superiority theory of humour is that people find it funny when they recognise their superiority over someone else, and this can include your past self. But all the other points still hold: not all humour can be explained in terms of status (eg. nonsequitur, some forms of wordplay), and it doesn't do a good job of explaining why we find the things funny that we do

Crossposts: http://erratio.livejournal.com/131311.html

So I've been reading Inside Jokes the past few days, and one thing that I find both interesting and alarming in their overview of cognitive theories is just how much each theory relies on contemporary technology, not just as a framework but in a way that indicates that the authors of each theory were mentally constrained by what was available. For example, the release theory of humour, the idea that humour is a release of nervous energy/tension, relies heavily on a gasoline model of cognition, where over time cognitive energy can 'build up' over time in the hypothetical pipes of our brains. And the frame/script model of cognition, where we have a bunch of scripts pre-built from common features of our previous experiences, is considered an example of 'just-in-case' processing, a kind of model processing that was widely in practice at the time. The alternative that Dennet et al are proposing? Just-in-time spreading activation, which is a modelling process borrowed from current economics and which is used extensively in inventory management in large companies. And of course our current models of cognition involve computation and neural nets and so forth.

All of this of course, points to the idea that part of the reason we've been having so much trouble with computational modelling of cognition is that computation might be the wrong metaphor. More wrong than the pipes-and-fuel model, or the gears-and-cogs models of the past? Probably not, considering how much more stuff we can do with computation in general. But I definitely wouldn't rule out another paradigm shift or two before we hit on a sufficiently accurate model of cognition that we can actually do stuff with it.

Crossposts: http://erratio.livejournal.com/130772.html

All the theories of humour I've seen can be described as falling into one of 4 categories:

1. Status/signalling
2. Incongruity/unexpectedness
3. Norm violation (has overlap with both the previous categories but has been advanced in enough different forms to deserve its own category)
4. Humour/laughter is a defense mechanism against negative emotions

I'll cover each one in a separate post. Since it's Thanksgiving break my aim for this week is to get a post out every day or two.

Status/signalling theories

According to these theories, the main purpose of humour is to signal how clever you are (so that more people want to mate with you), and in the process raise your status relative to everyone else (so that more people want to mate with you). So if you make a joke and people laugh, you've scored points against the people laughing (since they've accepted this particular bid for dominance) and against whoever was the butt of the joke. If a person or group is the butt of the joke you've also increased ingroup-outgroup bias, by explicitly pointing out one of the differences between us and them and implying that we would never be as dumb as them.

What this theory is good at:
It does an excellent job of explaining why we laugh more at jokes told by people with high status than low and why other people's laughter makes things funnier. It explains why a large portion of humour consists of clever insights and wordplay. It explains why people enjoy slapstick and mean humour (other people getting their status lowered), and it explains why obscure jokes and parodies are funnier to the people who get them (more ingroup effects, plus the general high status of knowing more about that field/derivative work). Oh and it also explains sarcastic/unhelpful humour, since not getting the joke means you don't have the same shared knowledge as the person telling it.

What this theory isn't good at:
It doesn't explain toilet humour, which is anything but clever, or why some jokes are funnier due to their unexpectedness while others are funny regardless of their expectedness value or get even funnier through anticipation. The 'mean humour' part of the explanation falls apart when you take into account people like my friend Anton, who's one of the least competitive/status-driven people I know and yet finds mean sitcoms hilarious, or my friend McKenzie, who describes herself as very status-driven but hates them. It also doesn't explain why only particular types of cleverness are eligible for humour - I've been taught by several quite brilliant teachers, and yet I somehow managed not to fall out of my seat laughing whenever they explained a new concept.

Trying to deal with the problems
Some of these objections can be dealt with given a suitable amount of wiggle room.

The variance in the :"mean humour'' category makes more sense if you introduce a term for empathy, where the default would be that mean humour is funny but to find it less funny the more empathetic you are. Then you could try running an experiment involving varying amounts of disassociation between the person being hurt and the person who's meant to find it funny. But regardless of the results, I think this still doesn't quite account for all the nuances of humour in "people getting hurt". For example, a guy getting hit in the balls is generally considered to be hilarious, even though this is one of the situations you would expect guys to have the most empathy. Other situations with comparable amounts of damage/pain, like getting punched in the face or a woman getting hit in the breasts aren't nearly as funny. Similarly, a person finding out their spouse has been sneaking out to pursue their dream of being a singer isn't funny, but a person finding our their spouse has been sneaking out to pursue their dream of being a clown is funnier, even though most people don't want to be either.

Toilet humour generally requires an appropriate context (ie. cleverness. Just blurting out "penis" isn't funny after the age of 15), and could be considered to have a status-related component in that only high-status individuals would risk talking about taboo topics. This doesn't feel like it explains the facts that well though - in some groups, toilet humour will get you a dirty look no matter how high status or clever you are. And the people who find toilet humour the funniest (children, some of the people at my old workplace) tend to be the ones with the lower status.

Types of cleverness: this isn't explainable at all without appeal to outside variables like unexpectedness, as far as I can tell.

Conclusion
Status explanations can't account for why we find the things funny that we do. But like pretty much everything else, humour has been co-opted into the status game and is treated as a marker for high status, which then feeds back into our sense of humour and influences how funny potentially-funny-things are to us.

Crossposts: http://erratio.livejournal.com/130046.html

I feel like procrastinating, so this is as good a time as any to start writing some of my half-assed research up. My plan of attack: Part 1 will cover the types of phenomena a good theory of humour would need to cover. Parts 2-5 (+/- 2) will cover current theories of humour and examine how well they account for different categories of humour. Subsequent parts will try to pull in other cogsci research and personal observations to make sense of it all. And I don't have a grand conclusion that I'm leading up to :)

Categories of humour that any model would need to account for:

* Ingroup humour.: jokes that aren't actually funny (assuming they ever were) but which people in your social circle use and which you and your friends use and laugh at anyway. (your mother and that's what she said tend to fall into this category)

* Obscure humour: the relatively well-known phenomenon where the more obscure the knowledge required to understand the joke, the more likely you are to find it hilarious

* Status effects on humour: you laugh more at jokes made by people who you consider to be higher status than you, and less at jokes made by people who are lower status

* Seeing/hearing other people laugh makes things funnier.

* Parodies and other heavily derivative humour are much funnier than you would reasonably expect.

* Jokes where the humour comes due to the unexpectedness of the punchline (absurdist humour falls under this a lot of the time)

* Jokes that are still funny or even funnier when you know the punchline in advance ("the Hammer is my penis"). At the extreme end of this are jokes that start out not-funny or marginally funny and become hilarious through repetition. (can't think of any examples of this at the moment)

* Puns and other wordplay

* Slapstick and other visual humour

* Mean humour (It's always sunny in Philadelphia is good for this. I can't stand it because the characters are all dicks but I know lots of people who love it)

* Toilet humour and other humour that derives from inappropriateness and taboo topics

* Humour that consists of saying things that are obviously false or unhelpful. This includes sarcasm, but there are also lots of instances of this that lack the connotations of sarcasm. (Example: one of my fellow grad students saying that preservatives are healthy and make you live longer, that's why they're called preservatives).

So, what have I missed?

Crossposts: http://erratio.livejournal.com/129739.html

So, I just attended a talk by Carol Fowler, an academic who works in articulatory/gestural phonology. She is awesome and her talk was awesome - I was more impressed by it than I've been by most other talks, even though is only varely related to my usual interests. I'm now going to try to capture as much of it as I can remember, and since it's about cogsci-ish stuff I'm putting it up here where other people with similar interests can read it, if they want. It's quite rambly though and I don't know enough about phonetics to really explain most of it clearly, so if you want the cool/easier to understand bits, skip down to the embodied cognition/mirror neuron section.

Relevant jargon:
gestures: when applied to speech production refer to movements of the tongue, jaw, lips, etc.
formants: the frequences of speech sounds. More specifically, most speech sounds show up on a spectrogram/graph/whatever as 2 or moer lines at various heights which our brain combines to produce a sound.
VOT - voice onset timing: I'm not really sure what this is since it's outside my area but it's one of the qualities of the speech signal that lets you distinguish between different sounds. VOT's vary across demographics in much the way you would expect, which is to say that it varies between individuals but there are also broad gender/cultural trends.
TMS: transcranial magnetic stimulation - it's gotten popular recently as a way to shut off areas of the brain but apparently it's also what they use for stimulating specific muscles directly.
Stop - a consonant that involves full stoppage of the vocal tract. The most consonanty consonants that exist. Letters like p,t,g are stops.

Overall argument: a really big part of language perception relies on embodied cognition type stuff, because trying to reconstruct actual sounds from a continuous stream is really really hard.

Phonetics stuff: Speech signals are not consistent. For example, /gi/ versus /gu/, the formants are what you would expect for the vowel parts (/i/ is high, /u/ is lower), but the onsets that make people hear the /g/ look like a little downtick on the first one and a little uptick on the second. The short noisy burst that is a stop looks almost exactly the same regardless of whether it's a /p/,/t/ or whatever, so Liberman guessed that speakers must be using articulatory information to disambiguate them. The proof of this can be seen in the McGurk effect, where you listen to one syllable and see another being mouthed and what you perceive is usually somewhere between the two but closer to what you see. The McGurk effect has been replicated in multiple modalities, including one Carol recounted one where the listener put their hand over her face while she mouthed various syllables, and another one where people got a puff of air on their necks to mimic the aspiration of the /p/ in /pa/ versus the non-aspiration in /ba/. Writing is one of the few modalities that doesn't show an effect. When you put other syllables such as /ar/ or /al/ in front of an ambiguous /pa/ or /ga/, people hear it in a way that indicates that they're overcompensating for the effects of coarticulation of the two consonants. (the other theory had to do with formants, but a Tamil linguist finally found a minimal pair to test both theories and it came out in favour of overcompensation)

The peception by synthesis argument: Someone (Liberman?) thinks people understand speech by modelling possible gestures until they find the correct one. Carol disagrees with this because in perception you only have a very limited time to work out what they said, and your brain doesn't enjoy being wrong because that means more work, so trial and error seems unlikely. Also, no one actually speaks identically so there's no way I can model what you said accurately in any case (although this one's kind of a weak argument, because it's a matter of getting close enough)

Mirror neurons and embodied cognition stuff: People primed with thoughts about old people moved more slowly on their way to the lift afterwards. Subjects made to hold a pen between their teeth, forcing their mouth into a 'smile', were more likely to perceive other faces as smiling. In speech perception, if people had a machine pulling their mouths up or down to mimic the shape of their mouth when forming various vowels were more likely to hear ambiguous vowels as the one corresponding to their mouth shape, even though they weren't making that mouth shape deliberately. Similarly, when TMS was used to stimulate subjects lips or tongue while hearing various consonants, they were more likely to perceive the consonant made with that part of the mouth. People watching other people walk have short bursts of neural activity corresponding to leg muscle movement, this kind of thing doesn't happen when watching a movement that isn't humanly possibly, like wagging a tail. Similar effects in speech perception. When TMS was used to knock out part of the articulatory apparatus, people's speech perception suffered. When subjects had their jaw moved in a specific way by a machine such that it didn't actually affect how they produced a specific vowel, it still had an effect on how that vowel was perceived by those subjects later.

The chinchilla experiment: chinchillas kept in a US lab were successfully taught to distinguish between /pa/ and /ba/. More interestingly, the acoustic properties they were picking up were specific to English - apparently the way English speakers distinguish between the two sounds (something to do with VOT's) is fairly unusual, most languages put the boundary somewhere slightly different. So that's evidence against there being a special human phonetic module for speech perception. Other animals can do it too, chinchillas are just the silliest and therefore one of the strongest examples of it not being anything special. But Carol mentioned that she's skeptical of this experiment and would like to see it replicated.

Questions: Do blind people have correspondingly worse speech perception since they lack a lot of the cross-modal information which is apparently so important? Studies of the mirror neuron/embodied cognition stuff in signed languages.

Crossposts: http://erratio.livejournal.com/129485.html

http://lesswrong.com/lw/dr/generalizing_from_one_example/4m0m

A couple of excerpts:

"It is dark and quiet in my mind all the time. Thoughts take the form of silently talking to myself. There are only words. No visual memory, no imagination -- I don't know what these things are, they are only words. Seeing things in the mind, hearing things, re-experiencing, exploring non-physical possibilities via imagination: these all sound like paranormal or supernatural experiences to me, literally, because what is normal and natural for me is the dark and quiet mind."

"Reading a novel or watching a movie is a lot like other things in my life: I am only engaging in the current moment of it mentally, the preceding parts are gone, because there is no way to hold on to them mentally. In order to watch a movie or read a novel, I make the effort the keep a running memorization going of a few key plot points in order to process the story. As soon as the movie is over or I've put the book down, it's basically gone from my consciousness, unless I try to think about it or talk about it, and then I only have access to those points that I memorized in order to keep up with the plot, which is a very bare-bones summary."

I consider myself to have extremely weak powers of visualisation compared to most, but the idea of being even more limited than that in all my mental senses kind of freaks me out.

Crossposts: http://erratio.livejournal.com/126807.html

Apparently, thought isn't as dependent on language as we might naively think

Excerpt:

"The man she would call, ‘Ildefonso,’ had figured out how to survive, in part by simply copying those around him, but he had no idea what language was. Schaller found that he observed people’s lips and mouth moving, unaware that they were making sound, unaware that there was sound, trying to figure out what was happening from the movements of the mouths. She felt that he was frustrated because he thought everyone else could figur e things out from looking at each others’ moving mouths.
.
In contrast to the absolute inability Idefenso had getting the idea of ‘idea,’ or his struggles with points in time, he clearly was capable of all sorts of tasks that suggest he was not mentally inert or completely vacant. He had survived into adulthood, crossed into the US, kept himself from being mowed down in traffic or starving to death. Moreover, he and other languageless individuals had apparently figured out ways to communicate without a shared language, which I find both phenomenally intriguing and difficult to even imagine (putting aside the definitional problem of distinguishing human communication from ‘language’ broadly construed).

Schaller highlights that learning language isolated Ildefonso from other languageless individuals. Schaller explains:

The only thing he said, which I think is fascinating and raises more questions than answers, is that he used to be able to talk to his other languageless friends. They found each other over the years. He said to me, “I think differently. I can’t remember how I thought.” I think that’s phenomenal!

"

That last part about not being able to remember how to think or talk to his languageless friends echoes other research that language is important for encoding memories (I don't have the link handy, but in short: young childrens' ability to remember stuff was shown to be strongly correlated with their progress in language acquisition). But the way Ildefonso is described above makes me think that the lack of ability to remember pre-language events might not be due to an inability to encode memories in the absence of linguistic symbols, but a result of not sharing enough mental context with your prelinguistic self to be able to retrieve them.