Daniel Dennett: The Evolution of Reasons

I’m here in Science Theatre 140 at the University of Calgary with Pliny and Daniel. Waiting for Dan Dennett to speak. He was to have spoken yesterday, but was delayed by aircraft mechanical difficulties.

4:17 pm. Dennett has just arrived.

4:31 pm.people settling into place. Dennett now being introduced. Tumultuous applause. Dennett starts by commenting that Marx says Darwin drove teleology from biology.

But did Darwin actually accomplish this? NS does not reduce telos to material and formal causes.

Two sorts of why questions:
How come?
What for?

‘Why’ is ambiguous between ‘how come’ and ‘what for’? This marks the difference between cause and intent.

Some biological questions have a what for answer. Physics questions typically don’t.
‘How come’ questions seek process narratives. All why questions have a how come answer, but only a few have what for answers.

Behaviourists often advert only to how come answers, no what for answers.

‘What for’ questions are about what is good. So hence normativity and hence the need for a process of correction. There’s two sorts of normativity.

Pittsburgh normativity: use of the ‘space of reasons’ to justify action.

Consumer Reports normativity: value and function.

The difference: Pittsburgh normativity corrects what is naughty, CR corrects what is stupid.

‘how come’ answers are typically failure analyses or historical accounts.

Natural selection is a algorithmic process neutral to the substrate on which it acts. But how does natural selection get started? The clue lies in cycles in abiotic life. There are many of these: seasons, days, tides, water evaporation and thousands of chemical cycles. Think of these as ‘do loops’. Cycles return to starting point after achieving something. And this may raise or lower the probability of something new happening. Over a billion years or so, algorithmicity increases. And so does parallel processing. Some entities will be more likely to persist than others.

In some cases algorithms may clobber other algorithms or may work with them serendipitously. Membranes can prevent clobbering. Source: Dennis Bray, Wetware. Mass production becomes mass reproduction. Details here are necessarily fuzzy. Reproduction is the creation of tokens of types.

Once tokens of types occur, they can explore different parts of the world. Once structures such as membranes appear, ‘why’ questions are equivocal: membranes have both a history and a reason for continuing to persist. This marks a gradual shift from different persistence to differential reproduction.

The minimum requirements are a metabolic system, genetic system, and a membrane. But is senseless to ask which one came first.

The Birth of Reasons

Once the first bacterium appears, we can look for reasons for the structure of the bacterium. Dawkins is skeptical about seeing ‘real’ design in biology. Apparent design is not real design, as seen in cartoons, eg.

But evolved stuff is real design because it works. So we should agree with creationists that nature shows design, but that there is no designer. NS is an automatic reason-finder. NS tracks reasons, creating entities that have purposes but don’t need to know them. NS itself doesn’t need to know what it is doing.

Dennett compares Darwin’s ‘strange inversion of reasoning’ with Turing’s use of humans to solve algorithms. This yields competence without comprehension. On this account, understanding is the effect, not the cause of the mind.

A lot of ‘sorta’ comprehension may add up to real intelligence.

Lobster traps and caddis larvae food traps both have reasons for their design but only the lobster trap reasons are represented anywhere. The termite castle is built bottom-up, without direction, competence without comprehension, unlike a cathedral’s designer.

Next question. Humans have reasons, but do animals? The biotic world is saturated with reasons, ‘free floating rationales.’

There is a reason why cuckoo chicks push other eggs out of their nest. But the chick need not know the reason. But when an ape does something for a reason, is it more like a human architect or like a termite colony? Apes may be natural psychologists, but they never get to compare notes.

Leonard eisenberg,s great tree of life.

Our natural tendency to interpret all design as top-down is anachronistic. It’s just wrong to say’in the beginning was the word’ since words are relatively new.

Wilfred Sellars distinguishes between the manifest image of the world and the scientific image. We represent reasons, but how is this explained evolutionarily? We naturally take the intentional stance, to interpret actions as actions of an agent. So do other animals, but we alone represent this stance via language, making it visible. And we over applied the intentional stance to non-agents. Rivers want to go to the sea, and so forth. Skinner condemns ‘anthropomorphizing humans.’

So we may have gone too far getting rid of the intentional stance. We should instead recognize reasons and design as really existing in nature.

Talk ends at 5:53.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s