Lecture 8: Thinking computers and free will Flashcards
open vraag gaat wss over materialism, functionalism, multiple realizibility en reduction
dus goed leren
wat is de status quo over mind-brain
individual mental states within a person at a certain time do correspond to brain states in that person at that specific time. however, due to mental realizability we cannot define these via the type-type identity theory, because this blocks the bridge laws. it is not possible to reduce laws/theories at the lower level.
= non-reductive materialism
wat is reduction
wanneer je een higher level theory (ideal gas law -> if temperature is increased then pressure increases) uitlegd door middel van een lower level theory (statistical mechanics -> if particles move faster they will have higher force).
er zijn niet veel voorbeelden waar reduction niet werkt
oke
wat is een vb waarin het niet werkt
meteoroloy - higher level
geology - lower level
je kan niet predicten van wat voor soort steen (graniet, volcano, sandstone) een climate zone opgemaakt is (one side could be wet, one side could be dry)
wat is een voorbeeld van reduction dat niet werkt in psychologie
tought of death -> emotions of fear
maar dit valt niet terug te leiden naar lower order neurology: want de tought of death kan komen door neuro pattern in cortex op elk punt, en de emotions of fear is een verzameling van increased heart rate, sweaty hands etc. dus het is niet 1:1, en er zijn ook individual differences.
materialists would describe that…
every mental state is related to one brainstate at a certain moment.
wat is de weak form van identity theory
token-token (er is altijd iets verschillends bij verschillende personen)
wat zou je zien bij type-type identity
allemaal dezelfde brainstates bij een mental state (thoughts about death = brainstate f)
multiple realizability blocks….
reduction & type-type identity
for reduction, … identity is not enough, you need … identity
token-token is not enough
you need type-type identity
waarom is token-token niet genoeg voor reduction?
This is because the lower-order collection of realizations (“sandstone or granite or volcano or …”) from the lower-order perspective is heterogeneous and arbitrary; can only be “grouped” from the higher order perspective. that is why people who believe in multiple realizability find reduction unlikely
- But note that reduction is an incredibly great achievement that is also rare in exact sciences
- Without it, brain science is just as interesting!
- Specific mechanisms at the lower level can be informative for the higher level, without reducing that higher level
- Because mental states are realized differently across people and across time, it is more helpful to define them by their role in the system than by the brain state they are realized in, which brings us to functionalism
oke
John Bickle argues that reduction is possible, but it changes the meaning of the theoretical concepts
oke
functionalism vond…
we shouldnt really think of mental states as the brain states, but rather the role that they play.
if you believe in multiple realizability, you also believe that….
why not realise it in something else than a brain?
-> turing machine
wat liet de turing machine zien
if these thought processes are essentially computer programs, then we should also be able to implement a mind in a computer/robot. -> the brain is thus not essential, as long as the robot can realise the right programs, then this robot also thinks!!
opens up the question: is consciousness really limited to us?
if a turing machine can think a little bit…
then it is just a matter of building a very good turing machine that will think as well as we do.
turing test=
what would we say if we had a computer that behaves indistinguishable from a human being? -> in that case you have to grant the computer consciousness.
but: from what moment on should we say that the case?
test: you are communicating with two people, a person and a computer, and if we do not distinguish between computer and human, than we should grant that computer consciousness (would be unfair and discriminating). turing maakt hier dus een soort ethisch punt: waar heeft de robot consciousness?
what is the difficulty van the turing test
how long can we deny a laptop consciousness? is it fair to not give robots this? Bv als riet nu zou vertellen dat ze een robot is, dan zouden we niet mogen discrimineren?
as soon as we stop denying electronics consciousness, there are a lot of implications
betalen voor services die ze doen, etc
wat gebeurde er in de 70s-80s
- the thinking computer is coming!
- multiple realizability: we can transfer consciousness from one medium to another -> we could transfer our own minds to things.
- so we also become immortal
who demolished this point of view (of being immortal)
John Searle
wat deed searle:
The Chinese Room:
he wanted to show that the turing test is too weak.
a thought experiment, in which you are literally a Turing machine. you are in a room where everyone speaks chinese and you dont. you receive notes, and you have a book that tells you what to respond via symbols. the people outside might think that they are corresponding with someone who speaks chinese, but you actually dont. -> here we do meet the turing test, but we would not ever conclude that the person speaks chinese. the computer does not actually understands us, there is no meaning at all. therefore this computer should not have consciousness.
verschil weak version en strong version van AI
- weak; we kunnen er van leren
- strong: actually consciousness
wat zei searle dus over de turing test
the turing test may be necessary, but it is not sufficient ground for consciousness.
he thought a computer could never develop meaning -> it is a useful tool for understanding human cognition, but only biological organisms can have consciousness.
wat gebeurde er door searle zijn publicatie
the strong ai thesis failed, because searle claimed consciousness is an essential biological phenomenon. the functionalists are furious: “we shouldnt ask whether the person in the room speaks chinese, but whether the system as a whole does -> the whole room with the person and the paper does read chinese? maybe it is the system as a whole that understands chinese” hier reageerde searle weer op met dat de theory echt implausibel is. hij zag functionalists als een ideology (2 verschillende subsystems voor engels en chinees). hij dacht ook dat computers geen intenties konden krijgen, en daardoor niet dingen kunnen wensen, en daardoor geen subjectieve experiences.
hoe is het debat met searle nu relevant
moeten we chatgpt consciousness geven? of is het een simulatie (zoals searle zei)
will AI ever be able to have intentions? can it wish for things? can it have emotions?
qualia=
subjective, individual experiences. (hard to make sense of what these are exactly, why we have them and how they relate to our material description of the world)
suppose that Fred is discovered: has better colour recognition than anyone else.
Freds optical system is able to sort out the colours on the spectrum, there are optical differences in the brain (and therefore there are relations in the material).
waar was het fred verhaal een soort metafoor voor
qualia; even though we know what happens in the brain, what is material, we do not know what fred’s subjective experience is like
mary the colour scientists
- mary born in black and white room, never sees colour
- learns everything there is to know about color perception
- will she learn anything new when she goes outside for the first time?
wat als je ja antwoord op de vraag van mary
this means that Mary’s physical knowledge does not include everything. therefore knowledge about physicala processes is therefore not the same as knowledge about subjective experience.
= materialism has a hard time explaining qualia
thomas nagel zei…
what is it like to be a bat: wonders whether it is possible to learn from objective descriptions from ‘what it is like to be a bat’ and see the world through sound. he concludes that this is impossible, and therefore subjective experience escapes the grasp of science
these thought experiments pose a problem for….
functionalists & materialists
functionalism & computers=
a turing machine can do operations similar to logical reasoning, but qualia?
materialism vs dualism (in terms of qualia)
if all there is is physical, why do qualia escape a physical description of the world
neither a … nor a … description of the mind seems to capture qualia
functional & material
dus welke stroming kan wel qualia uitleggen
dualism (er is gewoon een nonphysical substance die subjectieve experiences veroorzaakt)
the hard problem =
many people think the brain produces experiences: “if a certain brain process occurs, then there is a subjective experience, otherwise there is not”.
but it is not clear how this existence of subjective experiences can follow from theories about the physical reality
(how is subjective experience possible at all? why do the physical properties that we have, come with subjective experiences?)
wat zei chalmers
consciousness does not really fit in the picture of scientific explanations between biology, chemistry and psychology.
free will probleem
the body has to obey the laws of physics, but then where do you make space for the mind to have an influence?
if the body behaves like a machine where physical causes have physical effects, how can we make room for free will?
wat is het lastige aan vrije wil, wat zijn de consequenties als bijvoorbeeld blijkt dat we dat niet zouden hebben
we are used to holding people responsible for their behaviour, which requires some sort of freedom. if that would not exist, our whole justice system would be based on nothing
we normally speak of free will when….
- the intention of behaviour precedes behaviour
- the behaviour was not necessary (you could have done something else)
- the intention of behaviour was the cause of the behaviour
hoe kijkt materialism naar deze free will
die 3 criteria werken niet goed door materialism.
william james=
my first act in free will shall be that i believe in free will
skinner =
did not believe in free will
wat voor experiment over intention and behaviour (free will)
Libet: investigated the timing of decisions versus behaviour
empirical studies seem to show that behaviour is already started in the brain before the decision is made
brain activity (=physical activation) well before they said they were aware of their decision. ver voor dat ze het zeiden of zagen etc. -> already activate behaviour before your decision.
wat zei libet
he did not deny free will, he said we prepare for many different actions, and free will is the decision to inhibit some of these.
determinism=
if you would know all of the particles, the direction that they went to and the speed etc, at one time point you can fully predict the physical state at t+1
states that all events, including human action, are ultimately determined by causes regarded as external to the will. Some philosophers have taken determinism to imply that individual human beings have no free will and cannot be held morally responsible for their actions.
dus wat impliceert determinism
there is no free will
is absence of determinism enough?
- a strategy to regain freedom is to deny determinism
- but is ‘probability of behaviour’ enough?
- exercising your will also presupposes determinism
- if my intention only determines my behaviour with some probability, i dont experience that as free will!
- does determinism contradict free will?
there is always chance, instead of necessarily one new state.
wat is de hele conclusie van dit verhaal
mental states are not independent causes. the real cause of your behavior is a physical state of your brain, which is just stuck in a causal chain
what are ways out of our free will issue
- criticise Libets experiment (increase om een andere reden)
- maintain that physics does not fully describe the world
- become a compatibilitst (both free will and determinism)
- conceptualize free will differently (not as a cause)
free will is now often considered as an … but it may also be …
explanatory entity
a phenomenon to be explained
phenomenon to be explained =
we differ in autonomy from cars and stones, therefore we call this phenomenon free will. question = how does that difference come about (difference in autonomy)
conclusie
- no one really understands how the mind relates to the body
- subjective experiences is a mystery
- demand for wanting free will puts these things on an edge
- misschien witnissen wij wel de breakthrough!