Brain Science Podcast discussion

"Brain Drain: Neuroscience wants to be the answer to everything. It isn’t" (Roger Scruton).


"There are many reasons for believing the brain is the seat of consciousness. Damage to the brain disrupts our mental processes; specific parts of the brain seem connected to specific mental capacities; and the nervous system, to which we owe movement, perception, sensation and bodily awareness, is a tangled mass of pathways, all of which end in the brain. This much was obvious to Hippocrates. Even Descartes, who believed in a radical divide between soul and body, acknowledged the special role of the brain in tying them together."

Read the full article here:

http://www.spectator.co.uk/print/essa...

From the Brain Drain article/link cited:

"Brain imaging won’t help you to analyse Bach’s Art of Fugue or to interpret King Lear any more than it will unravel the concept of legal responsibility or deliver a proof of Goldbach’s conjecture; it won’t help you to understand the concept of God or to evaluate the proofs for His existence, nor will it show you why justice is a virtue and cowardice a vice."

That's true, certainly for the foreseeable future. What it might do is help us understand why and how humans are capable of having an emotional response to music, literature, dance, or visual art in the first place.

"As though we didn’t know already that people feel pleasure in grooming and co-operating, and as though it adds anything to say that their brains are ‘wired’ to this effect, or that ‘neurochemicals’ might possibly be involved in producing it. This is pseudoscience of the first order, and owes what scant plausibility it possesses to the fact that it simply repeats the matter that it fails to explain. It perfectly illustrates the prevailing academic disorder, which is the loss of questions."

A completely unnecessary, obfuscatory argument and quite the contrary; humans seek out these behaviors like music and art, which first appeared in early humans as religious rites. It is the very fact that the brain's reward centers 'light up' when engaging in them that demands explanation. Of course we know it's due to wiring and neurochemicals, but why should this be? It is incomprehensible that we can take these ritual behaviors for granted, that they just happen to exist, and there's no more need to ask and pursue. No other animal that we know of has these ritual behaviors, at least not to the degree humans do, so what's the difference between humans and other animals? Brain scans and neuroscience, in general, are a portion of the investigation, as are psychology, anthropology, etc.

The real issue here is what is your agenda? My agenda is to force all human behaviors into the constraints of biological evolution. You reject some science as inappropriate for studying the human condition, but why does it have to be one or the other? Why can't it be both or all?

Yes, understanding consciousness, or even the human condition, remains an intractable problem, but are we to simply give up using science to elucidate it? Is there no more knowledge to be gained from neuroscience? Since the earth is the center of the universe, there's no more need to look at the cosmos, right?

It’s a rather silly, archaic article, attempting to put a case for the humanities as a kind of uber-science, and its criticism of the findings of neuroscience as being in some way tautological is quite absurd.

Of course its important if pair-bonding and relationships with children result in rises of oxytocin, just as the advice that putting on the bathroom light when you go for a pee in the middle of the night, lowers melatonin and leads to insomnia. No doubt the humanities advocate would lie there mulling over the problems of the day, rather than getting back to sleep, helpless to change his situation.

Either it is under the control of his own conscious will, or it is some act of God. That’s the dichotomy that his way of thinking results in.

The example he gives of a pixellated image is just silly. The real wonder is that there are some 40 different areas in the occipital lobe, that represent different aspects of the image, all having some regular mapping onto the retinal image, that in some way currently unknown, are merged to form our conscious image of our surroundings.

And that when we see a fast-moving object in our peripheral vision where the retinal cone density is small, and so is the neuronal density in the occipital lobe, we unconsciously postulate some recognition of the object.

When you are driving along at 30 mph. with parked cars either side, you are far more likely to identify a child or a pet, than you are if you are on an open road in the country at 60 mph. And that identification governs whether you swerve or brake or instead let the object take its own chances but concentrate on minimising your own chances of losing control of your vehicle.

So neuroscience is giving us the essential scientific information allowing us to take a stance here on questions of culpability.

Another really interesting finding is that we are all hypocrites, quite routinely. Cognitive Psychology experiments have shown that often the stimulus for our response occurs after the response arose, and yet the experimental subject will swear blind that it was that stimulus that caused the response. All right, that isn’t yet Neuroscience, but fMRI is being applied to detect the brain areas that respond to the stimulus, and those that initiate the response, and some day the data movement between the two areas might also be imaged.

These are techniques that will some day allow us to test the a-posteriori rationalisation in the unconscious minds of witnesses.

There was a recent TV science program that showed subjects watching a video of a suspicious looking person darting out of an alley, carrying something in his hand, with the sound track showing that one observer at the scene shouted to another "He's got a shoe". All the subjects reported that the warning was instead "He's going to shoot". To suggest that these two conflicting messages were in the conscious mind of the hearer, and that he arbitrarily chose one that suited his own prejudices, is malicious. Our understanding of speech uses both bottom-up and top-down reasoning, and the phonological loop holds only 5 +/- 1 words or higher-level concepts at a time, before rejecting all the alternatives but one, so that these can never later be retrieved into consciousness.

The humanities, as Scruton argues, would pretend that these alternatives were all "thoughts" in the conscious mind of the person standing before you.

Scruton also thinks that cowardice is a vice. I would argue that it often gives an evolutionary advantage, whilst bravery is often a quality that is whipped up by old armchair generals to persuade more gullible young people to endanger their lives.

We used to train our young people in boxing, before scientific evidence showed that it caused brain damage. The bad old days before the ascendance of scientific proof could be pretty awful, even if science has given us more hideous weapons than before.

In the UK, our writers and journalists are in general terrifyingly inadequate in their knowledge of science. When the Fukushima nuclear disaster occurred, the Washington post published schematics of the inside of the reactors within a few days. We had to wait some 6 weeks before our Guardian newspaper provided these, and our TV just made reassuring noises all through the developing disaster.

This reliance on the humanities as a kind of uber-science can be very dangerous indeed. Logic is
interesting and often helpful, but statistics applied to a theory of knowledge is usually necessary
to provide a complete model.

The more I think about Scruton’s article, the more I think it is dangerously wrong.

Chomsky would presumably meet with Scruton’s approval, as being a humanities expert. Chomsky’s model of language understanding contained a lot of well worked-out ideas, like the concept of phrase structure grammar. But it also had a lot of disprovable ideas, like the distinction between complements and adjuncts, which was apparently thought up to show the similarity between nouns and verbs in the way they bind with the words and phrases that they govern.

And the biggest error was the idea of transformational grammar, where some sentences have an underlying more simplistic alternative grammar, in contrast to their surface form, and have been changed, for example into a passive form, by reversible transformation rules.

Almost certainly this does explain the evolution of grammar as it developed over millennia by interaction between speakers, but psycholinguistic experiments have clearly shown that passive sentences take no longer to understand than active ones, which is inconsistent with the application of transformation rules.

The theory has led to frighteningly complex models like Head Driven Phrase Structure Grammar. In the domain of Artificial Intelligence, Charniak published a book in 1993 making a counter-argument, and warning that such grammatical approaches would not deliver practical results, and that statistical approaches would be better.

Work by psychologists in the 1990s demonstrated that there are multiple dictionaries in the brain, to deal variously with speech and text, input and output, and whole-words as opposed to phonemes and graphemes. A book by David Caplan in 1996 explained these models of language, and they are also explained in a book by Stanislas Dehaene recently treated on the Brain Science Podcast.

Where there are multiple systems with functionality that overlaps, differing decisions must be reconciled by some kind of voting strategy. Votes have to be independently weighted according to the confidence in them, so we get rankings of votes by their weightings.

Confidence and rankings are the objectives of statistics. Pinker makes a clear distinction between rule-based reasoning and statistical reasoning. He describes how children initially learn the past participles of verbs by wrote, which can be implemented in a neural net, and later generalise the rule that –ed should be added. He has tried to localise the two approaches by functional MRI, but unfortunately has been unsuccessful to date.

It does appear that we need models that reconcile the rule-based approach with the statistical one.
Jackendoff appears to be moving in this direction with his recent books. These treat a multiple-level model of language, with levels for morphology, grammar, semantics and ontological knowledge. Information is exchanged both up and down the levels.

The representation of ontological knowledge is also rather unclear . The classical approach worked-out by people like Sowa is heavily logic-based, and results in deep hierarchies. We can download Princeton’s Wordnet, and play with a very deep hierarchy of this sort. But it will be clear to us after a few hours that quite arbitrary classification decisions have been made, making the model very incomplete.

Work by the psychologists, such as “The Big Book of Concepts” by Murphy, shows that “folk-taxonomies” contain typically only 3 levels, and use prototypes which are typical examples of a class. Since educated specialists at least pay lip-service to the deeper models, we must in our brains have at least two models, and probably more, and weighting and ranking must be being used to reconcile them.

It seems to me that Scruton’s viewpoint is that we should not even try to reconcile these different approaches, and should leave discussions in these areas to specialists in the humanities.
That seems to me to be a council of despair.

John,

Good stuff, but your argument uses logic and science. It's a paradigm that many use selectively, sparingly, or not at all. Scruton's paradigm or agenda is not fully clear. It could be religious but more likely it's a form of secular humanism. The goal in many cases is to oppose determinism and reductionism in order to preserve free will and the human potential to overcome the tyranny of their biology. The progressive march of science increasingly suggests this is not possible, certainly in near evolutionary timeframes.

People are dedicated to their beliefs at all costs. The work of Kahneman, Tversky, and many others shows how people arrive at decisions and understanding through heuristics—"common sense," rules of thumb, and cognitive biases. Scientists have the ironic challenge of suppressing their own prejudices in order to create hypotheses and interpret experimental results, which they do only slightly better than non-scientists, maybe, sometimes. Most people have no such encumbrances and adhere tenaciously to their acquired beliefs.

Bottom line: your elegant arguments are no match for deeply-rooted biases. But don't stop the dialog in any case.

I am not usually sympathetic to Roger Scruton but I am astonished by the lack of skepticism about neuroscience shown in most of these responses, as well as by the apparent belief that neuroscience can tell us what it means to be a person.

There is a difference between using scientific knowledge to produce, say, an electric battery, or antibiotics. Science can provide information that enables us to produce certain types of results.

But using science to answer big questions about the world - where did the universe come from for example - is entirely different. There have been many attempts to answer that question but no one would suggest that science has the answer. There are some answers but they may be entirely wrong. They will certainly be altered and improved upon.

Many, if not most, of the claims being currently made about brains and minds assume answers to big questions about the nature of the mind have been answered. The word "claim" is important.I am not suggesting we should not be asking questions and suggesting answers to clearly articulated problems.

In these comments there is no acknowledgment of the inability of neuroscience to tell us much if anything about the mind, nor any acknowledgement of the limits of FMRI and the necessary limits of laboratory research generally into humans.

Scruton uses the concept of the homunculus. I understand this as primarily an issue of agency. Unless or until we understand the relationship between our brains and our minds, rather than the inventive correlations of FMRI, then we should not be implying the brain has agency in the sense that a person does - any more than we would suggest our knee has agency.

I also agree with Scruton that claims that understanding of the brain has removed agency are incorrect.

None of the examples given above address this point. They assume that we can treat knowledge about brains as knowledge about minds and persons.

Aside from that, I think Scruton is mainly concerned to attack others in the Humanities who use neuroscience in ways that he does not believe are valid. I am surprised to find myself agreeing with him but I certainly find the use of neuroscience in the humanities often exposes the great limits of those using it and of neuroscience.

The notion that scientists successfully suppress their biases is fantasy. I recommend Cordelia Fine's Delusions of Gender for a description of much bad (neuro)science. Even if you think you are going to disagree or you are sure she is just a feminist - if you care about science give it a go.

And in passing, I have also found that knowledge of science in the UK is lower than in the USA and also in other Western cultures I have lived in.

Hera wrote: "I am not usually sympathetic to Roger Scruton but I am astonished by the lack of skepticism about neuroscience shown in most of these responses, as well as by the apparent belief that neuroscience ..."

I think you will also enjoy the book Robert Burton has coming out in April 2013 called A Skeptic's Guide to the Mind: What Neuroscience Can and Cannot Tell Us About Ourselves. If all goes as plans he will be featured in the April episode of the Brain Science Podcast.