Categories
Week 5

Crossover in biology and computer science

Last week, Rosie was talking about how biologists are uncertain about the benefits provided by sexual recombination. Particularly, she said something about how ‘the numbers’ don’t show any benefit to the organism for sexual recombination. If I recall correctly, sexual recombination is crossover (that’s how I’m going to treat it for this post and someone needs to correct me if I’m wrong).

Then, in my computer science class today, we were talking about Stochastic Local Search algorithms, of which Genetic Algorithms are a subset, and the lecture presented the introduction of crossover as the defining factor of GAs. The impression that I got was that, in computer science, crossover was assumed to be beneficial. And since computer scientists are generally good at calculating the generalized efficiency of algorithms, I assumed that there must be some fairly definite benefit provided by crossover, which contradicts what Rosie said about the mathematic benefits of biological crossover not adding up.

If GAs in computer science benefit from crossover and biological evolution doesn’t, then I see three possible outcomes. 1) The analogy between GAs and biology break down in this case – which is the default assumption. 2) The computer scientists haven’t done strict tests of crossover, and it will turn out to be unhelpful, as in biology. Or 3) GAs and biology both benefit from crossover, meaning that there will be mathematical models of the benefit of crossover from computer science which might be able to inform biology.

I have some discussion on these three possibilities. But I’m going to have to hold off for now.

Categories
Uncategorized

Mutation and Recombination

Rosie’s lecture got me thinking about my interest in the evolution of imagination, and whether it is useful to postulate ideas or memes as independent of the minds of thier hosts. Bascially, I want to test how far this metaphor can stretch, while remaining critical it might burst.

In genetics, mutations are changes within an individual, like typos in base pairs when DNA polymerase and other repair enzymes fail to spot and fix them. The rate of mutation tends to be very low, a balance between perfect copying and the physiological cost of preventing them.

Recombination of genes happens between (sexual) individuals and no one has yet been able to come up with a good theory of why it occurs when this process does more harm than good in mathematical modelling.

One of the main criticisms of the meme concept is that culture cannot be examined in bits or composed of independent units. But like religion, culture must be made up of something. Even genes do not “work” in isolation, so we should not expect memes to either. Genes used to be fuzzy and invisible too, before we could observe DNA and chromosomes; they are useful theoretical artifacts because when a sequence gets knocked out, it has some observable effect. But can a meme be knocked out and shed light on culture? Not likely.

So what would a meme be, physically? A sequence of neural connections or pattern of activity?

How would a meme mutate or recombine? Does it have fidelity, fecundity, and stability (longetivity)?

In Conceptural Integration Theory (CIT, also known as mental space theory), developed by Gilles Fauconnier and Mark Turner, a set of activated neuronal assemblies are momentarily marked off in a so called “mental space” (imagine a circle with a realistic painting of a woman as the neural pattern, composed of various parts such as head, neck, torso, color, shape, light, etc. and another circle with an African mask as the pattern, also composed of various parts such as head, shape, light, etc.), and the patterns in each space are run together in a simulated new “blended space” (imagine a new circle with Picasso’s famous painting Ladies of Avignon. This is a very crude example, but you get the idea.) Basically, the this theory looks vaguely like a venn diagram using concept maps. It has been critiqued for being fairly empirically untestable so far.

The mental spaces are short term, constructed via information stored in longer term memory networks and associations. In the example of the expression “digging one’s own financial grave,” there is the creation of two maps or “mental spaces” for the domains Grave Digging and Financial Failure, respectively. In each space, there are association networks such as Gravedigger, corpse, and death for “Grave Digging” and the unawareness of consequences, suffering, bad decisions for “Financial Failure” which are blended in a new space to make meaning of the metaphor (Slingerland 2008:178, 186). Blending allows us to conceive of “As if” scenarios, and build upon them, ideas upon ideas, memes spawning more memes.

What I want to ask, is whether this blending is similar to recombination or mutation.

A meme would be a neural pattern that codes for something very simple in a larger association network, such as the shape of a head or an eye, or changes in pitch or the identification and categorization of nouns, verbs, etc. A cluster of memes would code for all kinds of stylistic representations, images of familiar symbols, tunes and story plots. Blending creates new memes, which may be selected for or against depending on the frequency of other memes. For example, if my network of memes related to funerals included links to dark colors, grief, pain, images of dull skies, tears were to suddenly encounter links to funeral jokes, laughter, bright skies, and ritual dancing, there would be competition in terms of the strength of those connections, which links I activate more than others (weak links eventually become extinct), and these depend on the strength of the input signals which I recieve from the larger cultural meme pool. In that way, memes can be said to be independent from their hosts.

There is no blending without input into two or more “mental spaces”. Input must come from past experience and stimuli from other minds, from a memetic environment or community (like a gene pool?). These spaces (chromosomes?) “cross over” in some sense, and elements are thrown away in the process, but it is not reassembled between “two individual hosts” but in the same mind, which then can be communicated to another, etc. New ideas are often intentional, not mistakes in copying, so they don’t really “mutate” in the same sense.

Mutation happens when mistakes are made communicating between minds. It takes a long time to learn how to replicate a letter, to teach a child to write, and reproduce recognizable symbols with stability and fidelity. It takes an even longer time to copy a pattern for a fictional character, which includes all the past experiences, thoughts, feelings, actions, associations, friendships, events, yadda, yadda of that charcter (passed orally or textually, and is prone to mistakes, but there is some stability that there is agreement on someone named Anne of Green Gables or Harry Potter). Sounds may be misheard, but there is enough stability that we can trace those shifts most of the time. Instructions may be lost and the product or technology is left behind for reverse engineering. Nevertheless, there appears to be a certain rate of mutation that makes some memes more fit than others when phemotypically expressed in a cultural artefact, like some tunes that we can’t get rid of, or the ubiquitous smily face.

What does it mean for ideas to mutuate, recombine or blend? Is culture like an organism? Is this a useful metaphor?

Categories
Uncategorized

Nature Review on gene-culture interactions

Sorry, I totally got flooded with stuff this week… writing up my weekly blog post tonight!

But before I do so, just came across something while browsing latest Nature stuff (my Wednesday night tradition):

How culture shaped the human genome: bringing genetics and the human sciences together Laland et al. 2010 Nature Reviews Genetics

(check out Table 1 for a quick overview…)

Btw, humanities-related evolution papers are becoming fairly common in Nature Rev Genet and Trends in Ecology & Evolution (hereafter referred to as “TrEE“)

It’s really cool how you not only have several evolutionary systems going on in parallel, but also they interact with each other in both directions! Now that’s getting complicated…and thus, fun! =D

PS: A totally random thing just out today: Apparently people have actually been working on the physics and chemistry behind spiderwebs being so awesome at attracting dew drops…turns out it has to do with the spider silk nanofibre structure (Zheng et al. 2010 Nature).

Categories
language Week 4

Language/Orthography Relationship

I’m not sure if this actually makes sense, but it’s a thought that’s been kind of bouncing around in my head. Ashley’s presentation on symbiosis got me thinking about the relationship between language and writing.

First off, most writing systems originated as pictographic at one time or another: that is, they started out as rather rough drawings of whatever it was the symbol was to represent. This means that originally, the symbol had more to do with the idea behind the word than the word itself. Since they were not related to the sound of the word in their correspondent language, the symbols could not reflect affixes or other morphology, and therefore any interpretation would have been based entirely on connecting words and word order. This means that reading pictograms would have originally been much more of an art than the pretty straightforward “sound it out” method that we get away with today.

But the really important bit is that reading the pictograms aloud would have not made any sense because of the lack of morphology and proper formations. At best, it would have sounded stilted and forced.

What this means, in short, is that orthography would have had little to do with the language to which it was linked, simply because it was in no way phonetic. If written and spoken language were originally separate and then became merged, it looks a bit – not a lot, mind, but a bit – like a form of symbiosis.

Over time, the common thread of meaning between symbol and sound becomes overwhelming, and the writing takes on features of the spoken language; in the case of the Phoenician alphabet, a phonetic resemblance was required in order to make the symbols match the syntax. In the case of Chinese, none of the languages have much on the way of morphology, so it was enough to stick to more-or-less pictographic representations.

Eventually, the reverse also happens: people mispronounce words based on their spelling and that becomes a largely attested phenomenon in a particular dialect of the language as a whole.

The last point I wanted to make, going back on the “symbiosis” note, was that having a system of writing dramatically improves the chances a language has of a) remaining unchanged, since there’s a constant record to which debates may be referred, and b) spreading, since a language that has a writing system, being more solidified, will tend to resist change more than a language without one.

Categories
Week 4

computer game that teaches an understanding of evolutionary processes?

So I spent the weekend making a computer game with a bunch of complete strangers. We went from nothing to a finished game in 48 hours, with things like sleep deprivation abounding.

In any case, with computer games on the brain, I am wondering how one might create a fun game that teaches the basic processes of evolution. Not the mechanics and dirty work of getting it running (that’s what engineers are for!), but the concepts it would try to incorporate and the abilities it might try to teach.

Let us assume that we are targeting high-school aged students who haven’t yet taken high-school biology. It is difficult to get them involved in studying (because it’s ‘boring’), but if they are really dedicated to the things they find fun and socially acceptable (like games!). They might already have all the formal biology education they’ll ever get, and know Evolution as ‘one of those theories about how animals, y’know, like, change.’

From the topics that we’ve covered so far, which ones would you consider most important? These can be within the context of biology or any other application of evolutionary theory.

How would you frame this in terms of a game?

(I’ll write my own response to this at some point, but I want to hear what others think first.)

Categories
Uncategorized

Lateral Gene Transfer Animations

I hope to really improve my presentation skills before the MURC talks, but I admit I rushed making this one, since I did not feel I was an expert on the topic. I found these really neat animations that you can watch for review:

Categories
Uncategorized

Population Genetics Plaything

Okay, so here’s the place you can download the really simple population genetics simulation thingy I mentioned today: http://wps.prenhall.com/esm_freeman_evol_3/12/3315/848837.cw/index.html

Your different alleles (so types of a gene, remember) are labbeld A1 and A2. You can give their respective genotypes fitness values, the higher the fitness #, the more they benefit the organism carrying them. Everything else is pretty self-explanatory, except maybe the inbreeding coefficient, which is the probability f of an individual breeding with a relative (who, by definition, is carrying the same alleles).

Anyways, to see what they’re talking about in the Lynch paper, play around with the population size and fitness levels. I have a screenshot of what I did, but cannot for some reason get it to upload on photobucket…

Categories
Uncategorized

Brief response to Lynch

I can’t say that I enjoyed reading this article very much; all this jargon is beginning to irk me. And then I find my mind wandering off somewhere with more sunlight and trees. Modulation? What?

My understanding is this: evolutionary biology tends to invite broad speculation on the part of non-specialists who think they’re qualified to make assertions based on their high school biology classes. (Which is why I usually only discuss this stuff when I’m drunk–no one’s going to be bothered by my baseless, uninformed opinions.) Why? Because it’s easier and more fun to fit scientific theories into one’s pre-existing ideas than basing these assertions on, say, observable data. The latter tends to take more work, an open mind, and all sorts of nonsense like that. Since evolutionary biology encompasses many philosophically sensitive topics (mostly just our creation and the reasons why we are the way we are), and since billions of years of its history are unobservable due to our irritating lack of time-travelling technologies, all sorts of people (including myself) like to butt in and dole out our two cents. In fact, the longer I take this course, the more evident my indecent lack of knowledge becomes.

But this isn’t as blame-worthy as I’ve ended up making it sound. World views are powerful forces, and not likely to change quickly, excepting some extreme happenstance. So, as we saw with that depression article, scientists of every flavor are not immune to such biases. Essentially, it’s good to remember that deeming a trait as better/worse, adaptive/maladaptive is not very useful and only marginally scientific. I like what someone wrote earlier about complexity: how do we define that, anyway? Because we tend to think that we are the most complex organism out there, just because we have invented things like the can opener and disco. Isn’t simplicity usually more stable, anyway?

Categories
language Week 3

An Alternative to Tree Structures

One thing we brought up in historical linguistics was something of an alternative to tree diagrams. Trees are really nice for representing linear branchings and genetic relatedness, but have a hard time describing what exactly is similar between two languages; as such, some people in linguistics have proposed what’s called the “wave model” of linguistics. Here’s a link to the Wikipedia Article.

For those of you who don’t want to read it, I’ll summarize it here. Basically, you put down the names of the languages on paper and then draw lines around certain subsets of them. Each enclosed area represents a single innovation that sets those languages apart from the rest that you’re studying. The primary advantage to it is that it’s based in featural commonalities, so it’s really easy to see what exactly particular languages share. Disadvantages include the fact that they’re painstakingly difficult to draw and to read, and sometimes the person making them screws up and you get lines bleeding into each other… they can be a real mess.

That said, since they allow you to look at certain changes, if you took things like, say, Early Latin, Classical Latin, Late Latin, a couple dialects of Old French, a couple of Middle French, and a couple of Modern French, those lines are going to tell you what sort of changes happened when and how, simply by mapping the linguistic changes onto the fairly well-known historical record of the area and cultures.

Also, wave models can make it very easy to distinguish which features are “genetically” conditioned (that is, those that the language retained from it’s ancestor language) and those that are “areally” conditioned (those that the language assimilated from other nearby languages). All you really have to do is look at the features shared between most of the languages in its area and those that shared between most everything else in its family.

Of course, life’s rarely that simple, since languages move with the cultures that speak them and so often times it’s no easy task to determine why the language has the areal features it does. But I think the idea’s kind of a cool one.

Also, and this is just a side note, we should have a tag for culture.

Categories
Week 3

Very brief thoughts for this week, as my brain is off on a different planet due to some personal stuff, so I haven’t really been in the mindset for nice crunchy analytical thinking for the last few days. So I thought I’d share a point that came up in a conversation between Yana and me on the bus to Safeway last Thursday to go pick up drinks. We were discussing the relationship between drift and neutral evolution, as the two ideas are conflated. Drift, in population genetics, is specifically related to the frequency of alleles already established in a population, rather than dealing with novel mutations. We tend to think of evolution as being strictly about novel mutations, but different alleles of a gene are essentially established mutations within a population. In this sense, drift is basically a subset of neutral evolution – change without selection pressure.

Also completely unrelated, but I had a moment of nerdy glee when I was looking for resources for my proposal and found that Koerner has a book called ‘Is there a universal grammar of religion?’. Sadly it is currently checked out, but I’ll get my hands on it eventually!

Spam prevention powered by Akismet