Cogsci midterm

What do 00blank000blank and 000blank00000blank represent on a Turing machine

2+3 and 3+5

Describe the main idea/argument of the Batts, reading, the main reasons for reaching their conclusion, what findings or ideas they describe, and what the significance of those findings or ideas to be

Batts argues Brain lesions (damage to neurons) are increasingly being used in criminal courtrooms to argue that criminal defendants should not be found guilty of crimes, despite their actions having caused harm. Batts notes that this may be warranted in some cases, but that it also a cause for concern since there is not an obvious link between having a brain lesion and failing to know the difference between moral and immoral action. There's a lot of evidence that having a damaged architecture of your brain can lead to abnormal information processing in all sorts of ways, but the connection to this sort of knowing the difference between right and wrong is much harder to establish. The discussion in Batts' article highlights the potentially profound impact of neuroimaging evidence on juror perceptions and verdicts .Since brain scans seem objective, jurors might be swayed disproportionately by then, perhaps without fully grasping the nuanced relationship between brain anomalies and behavior.

Describe how Benjamin Libet's experiments provide reasons for thinking that our actions are not freely chosen.

Benjamin Lilibets experiments provide reasons for thinking that our actions are not freely chosen by showing example of how one could not freely move finger, which puts pressure on the neuroscientific challenge to the idea of free will, the idea of being able to do otherwise. His experiments basically show/provide that if our brains are ultimately responsible for our choices and there is nothing like free will located in the brain, then perhaps we do lack free will after all,, If our brains are ultimately in control, it would seem to give us pause in thinking that it's obviously the case that we can choose what to do In Libet's experiment, participants were asked to move their finger at a time of their choosing while staring at a dot that moved in regular patterns across a screen. Participants were wearing an EEG or electroencephalogram, which is a device that measures electrical activity in the brain. Participants were told, participants told the experimenters where the dot was on the screen when they first decided to move their finger, right? So they moved their finger, but then they report to the experimenters where the dot was on the screen, not when they moved their finger, but when they decided, okay, I'm gonna move my finger now, right? But the EEG picked up was something called a readiness potential or an increase in the brain's electrical activity associated with movement 550 milliseconds before the participants became aware of their decision to move their finger, right? So there's this idea here that, you know, time is moving along, the EEG is on their scalp, it's recording electrical pulses from the brain. before we even become aware that we intend to do something, our brains, it seems, are already getting ready to do that thing, You see a rise of the readiness potential right around here. It's only here that they become aware of their intention to move their finger and hear that they move their finger,right? So this idea that our brain, a strong interpretation of LeBetz results is that our brains know what we're going to do before we do it, This is a strong interpretation, and these results have been replicated repeatedly with longer gaps observed between the first detection of the readiness poten

Describe the main idea/argument of the Bisson reading, the main reasons for reaching their conclusion, what findings or ideas they describe, and what the significance of those findings or ideas to be

Bisson's argument/main idea is that the aliens (of the story)make the mistake of thinking that certain kinds of implementations make cognitive function impossible. They think that humans can't possibly have advanced cognitive capacities, because "meat" is such a strange way for these capacities to be implemented (weird idea of thinking meat). The aliens lack creativity about the wide variety of ways in which cognitive algorithms can be implemented. It's Related to implementation Marr level also Turing machines can be seen as implementing cognitive algorithms and meat doing the same isn't crazy. But aliens missed this point

Describe the main idea/argument of the Bucholtz and Faigman, reading, the main reasons for reaching their conclusion, what findings or ideas they describe, and what the significance of those findings or ideas to be

Buckholtz and Faigman argues that despite having considerable promise, neuroscience has failed so far to provide any good system for establishing whether a defendant was capable of the level of self-control necessary to be criminally liable. They highlight the group-to-individual ("G2i") problem. This is similar to the type-token distinction. Basically says that even when there is a general correlation between a brain abnormality and violent behavior, it is difficult to show that any violent individual's behavior was cause by a brain issue. group to individual problem. This is similar to the type token distinction. Even when there's a general correlation between,even when there's this general correlation between a brain or a normality and violent behavior, it's difficult to show that any violent individual's behavior was caused by a brain issue. We may know that in general that people with this brain lesion tend to have more violent behavior, but it's still hard to show for any single person that that problem with their brain caused that problem with their behavior. Colton Fagiman, argued that despite having considerable promise, neuroscience has failed so far to provide any good system for establishing whether a defendant was capable of the level of self -control necessary to be criminally liable.

Describe the main idea/argument of the Carandini, reading, the main reasons for reaching their conclusion, what findings or ideas they describe, and what the significance of those findings or ideas to be

Carandini argues that just understanding the algorithms run by neural circuits is not enough to understand cognition but that we also need to understand the goals of information-processing systems This relates to the computational level in Marr's levels of analysis The article also mentions two alternative approaches to Marr's bridge the theoretical gap between neurons and behavior: building a connectome (full circuit diagram of the brain) and building a simulome, a simulation of brain circuits. The simulome approach may be more tractable than the connectome approach because of exponential growth in the number of connections between neurons in the brain, attempting to map all of these connections in a connectome might be infeasible or difficult to accomplish in polynomial time. A simulome might avoid such a "combinatorial explosion" by simulating smaller sub-networks of neurons. Simulating such sub-networks might still yield useful understandings of canonical computations. The simulation approach is more "generative" - scientists can play around with the parameters of the simulations and observe computations emerge at the scale of a few neurons. However, the connectome does not allow scientists to observe emergent behavior, and may be too "big" to analyze effectively - this would be like attempting to understand the behavior of an LLM by analyzing a list of the billions of weights and biases of its neural network.

Describe Chalmers' "Zombie Argument" intending/purpoting to establish an "explanatory gap" between physical descriptions of the mind and phenomenal consciousness and describe a counter argument to Chalmers thought experiment

Chalmers idea is that there's something called P zombies or philosophical zombies to make the point that this is a distinctly sort of philosophical object a kind of thought experimentt and what Chalmers argues is that zombies are at least logically possible; There's no reason why they couldn't exist. . It is not a tautology That there are no zombies. Right saying that there are no zombies in the world is not the same as saying that two plus two equals four, right? Chalmers is clear to think he doesn't believe that there are any zombies out there in the world. He doesn't necessarily think that there are people who walk and talk and think like human beings, but are not human but don't have any phenomenal consciousness. He's just saying that there's no logical contradiction in supposing that there are What he concludes from this is that neurobiology can never explain why we have the conscious experience we do this is another version of this idea of the explanatory gap This idea that there's an explanatory gap is sometimes called the hard problem of consciousness This is this idea that there are easy problems of consciousness like you know sort of conscious people tend to do, when people tend to report more or less consciousness, then there's a hard problem of consciousness, which is why anything is conscious at all. Because the idea is you could do brain surgery, you could use fMRI, you could do all kinds of physical tests that you wanted on a zombie, The explanatory gap here is that neurobiology can never explain why we have the conscious experiences that we do The churchlands counter argue to this-What's physically possible? Not what can I conceive of, but what's actually the case and what could be the case out there? And what's the case is that everything we know of with a sufficiently complex brain that isn't impaired in some way as a human being, we think of as having conscious experience. And so that should be enough to convince us that beings with suitably complex brains have conscious experience. And whether we can imagine the existence of p zombies is neither here nor there in this argument about what consciousness is and where it comes from, right? Churchland will say, "Look, I don't know the full s

.Describe the main idea of the Coltheart reading, the main reasons for reaching their conclusion, what findings or ideas they describe, and what the significance of those findings or ideas to be

Coltheart gives an overview of the modularity thesis in cognitive science, highlighting the main features that Fodor thinks are typically of mental modules. Whereas Fodor doesn't see any of these features as necessary for modularity, (doesn't want to commit to saying there are conditions), Coltheart disagrees and thinks that that modules aren't necessarily domain-specific (being usable in a particular demain is what a mental module is ). Coltheart's main claim is that a necessary requisite for cognitive system to be considered modular is its domain-specificity. Block's view that "a cognitive module may be decomposed into a smaller set of modules" (Coltheart 118) is necessary for Coltheart's main claim because, if it wasn't true, that would mean that a cognitive system that is domain-specific (e.g., the spoken-language module) would not be considered a module because it has different levels of representation that would be considered sub-modules (the processing of lexemes and phonemes) - even though it has the same input/domain (spoken language). Hence, Block's view allows for interlevels of representation that are sub-modules within a module - which in turn allows Coltheart to broadly define modularity as a cognitive system that is domain-specific in its input but not necessarily in the representation of given input with a module. Colthart, on the other hand, does think domain specificity, as you read in the reading, that domain specificity is definitional of a module. That, for him, is sort of what it is to be a module.

Why is connectionism opposed to the modularity thesis?

Connectionism is opposed to the modularity thesis because rather than saying we have different cognitive modules doing different things we've got one neural net that's kind of able to do it all. It goes against modularity by putting forth that we actually have an architecture for doing this kind of inference or domain general way. We don't have to say that there are sort of separate modules for each thing. So the idea here is that these neural networks are more domain general than... That's one. Then faculties are meant to be. Faculties are meant to be specific. There's this idea that if the modularity thesis is to be true we're going to have to point to the places in the brain where the different modules are implemented, but it seems like things in the brain are much more flexible than that, which would potentially give us a reason not to support modularity. However, connectionist approaches have yet to provide a definitive rebuttal to the poverty of the stimulus argument, but that is not to say that such a rebuttal may not be forthcoming in the future, or it may indeed never occur

Describe in depth what connectionism is.

Connectionism is the idea in cognitive science that we can explain cognition in terms of neural networks. An important rival to the modularity approach. An archutrcture for inference rather than domains for each thing. Rather, than saying we have different cognitive modules doing different things connectionism finds we've got one neural net that's kind of able to do it all. IConnectionism, inspired by artificial neural networks,like chat gbt is the main empiricist challenge to nativist modularity approaches. So say likeAn extremely well-trained neural network implemented not in, of graphical processing units like chatGPT is, but instead implemented right in our brains. If looking at how a complicated neural network could be impleente in something as small as the brain we can so we actually have an architecture for doing this kind of inference or domain general way. We don't have to say that there are sort of separate modules for each thing. Rather, we've got one neural net that's kind of able to do it all. It's able to take in all kinds of different inputs and put out the right outputs, whether it's doing language or recognizing sentiment or doing verbal reasoning or doing mathematical reasoning, et cetera. The idea is you've got these connections, you've got these weights, you've got neurons connected to other neurons, you've got weights to the connections and all of that together determines the values taken by all the neurons.

Describe in depth the type-token distinction between causal claims (or, alternatively, the "group-to-individual problem") and its relevance to the use of neuroscientific evidence in criminal law.

Criminal law is about token level causation; And criminal law is about, did this defendant cause the crime?And when we talk about causation in the context of brain abnormalities in criminal defendants, the question is, did this defendant's lesion cause them to commit the crime. Rhere's this central problem for the use of neuroscience in the courtroom, and this is this problem of causation. in the criminal law, when we want to use neuroscientific evidence to say that this brain abnormality caused this person's criminal behavior, what we're trying to make is the token level claim about what happened in this person, this brain lesion, not the whole type of events called brain lesions, but this brain lesion. We have to make a causal judgment on whether their brain and not, you know, a deficiency in the architecture of their brain and their brain physiology, rather than intention to cause harm is responsible for their actually causing harm. So what is that causal change? Does it go from intentions to actions to harm? Or does it go from brain deficiencies to actions to harm? And so science is all about searching for causal patterns or trends, It's all about looking out at the world, looking at all the data and finding where the causal patterns are. And these causal trends are sometimes called type level causal trends in reference to the type token distinction. Even when we have good evidence for type level causal relations, it can be conceptually difficult to use them to explain token level statistical claims, even with a high degree of confidence. It's difficult to even to say that a patient's brain state made them more likely to commit the crime. It's just very hard for science to establish facts on that kind of token individual level. We're interested in this defendant and their brain lesion and their violent behavior. And this is the kind of thing that science does not do as well at establishing, typically. Yes, that's exactly right. Yeah, so the group to individual problem, another way of thinking about it is this type token distinction and causation where in philosophy of science, we think a lot about this distinction between type causation and token causation. there's this group to individual problem of we

Define and describe the six features that are typical of a mental module.

Define and describe the six features that are typical of a mental module. Human thinking can be understood by appealing to humans making use of different cognitive modules at different times. Jerrt Fodor describes the following six properties that mental/cognitive modules (basically later version of faculties) typically have, as a starting point for his analysis. Domain-specific mental faculties The idea that heres a task in the world it is specifically designed for These are faculties that we apply in a particular domain of inquiry. Usually sound like vertical faculties., could be hor When we say there's a specific domain in which memory gets applied, namely when we're remembering things, and that's the domain in which we use that faculty, Colthart, does think domain specificity, domain specificity is definitional of a module (fodor doesnt make definitions). That, for him, is sort of what it is to be a module. Innately specified (i.e., not learned), We are born with the ability to develop faculties Neuroscience focus* We don't assemblewithout architecture already provided. Looking at how this architecture is inherited physiologically. Hardwired (implemented in specific, localized neural systems), There is a place in the brain you can go and find it. Relates to the neuroscience * Cognitive modules are implemented in specific localized neural systems. So this could be regions of the brain or coordinations and connections between different regions of the brain that we can point to and say, these are the regions responsible for this domain-specific faculty, right, which this is, again, this idea of a cognitive module. In relation to Marr , You might think of the modules themselves as capable of running different algorithms, but we need to say where they're implemented. Part of being a module is typically implemented in a specific place that we can point to. Nativism, language hardwired into brain Autonomous (does not share horizontal resources with other systems) They do not share horizontal resources with other systems. A classic example of a horizontal resource was this idea of memory. The idea is if you're exercising one of your vertical modules, one of your vertical faculties, and it uses memory, that memory

Why does enforcing criminal law, in many cases, requires what psychologists call "mind reading."

Enforcing criminal law, in many cases, requires what psychologists call "mind reading."because our capacity for mind reading enables us to make inferences about how another person will behave under different conditions based on that model of how that person thinks, and mind reading is having a model of someone else's mind which is necessary for settling the question of whether anyone intends to do anything (like committing a crime). The question of whether someone intends to do something is central to settling the question of whether they could be guilty or not guilty of murder. Most serious of consequenes can depend on this capacity for mind reading, Mind reading is a psychology term rferrig to out capacity to infer others mental states. In general we attribute goals, beliefs, desires, etc. to other cognitive agents in our Environment enabling us to make inferences about how the person will behave under different conditions, based on our model of how that person thinks which is a ery useful capacity for a social Animal and necessary for settling the question of whether anyone intends to do anything

Describe the main idea/argument of the Espetein, reading, the main reasons for reaching their conclusion, what findings or ideas they describe, and what the significance of those findings or ideas to be

Epstein provides an objection to turing machine computational theory of mind ; that Our brains are not computers. Intelligent behavior requires being a part of the environment and interacting with it, and this is distinct from computing. For example, When a baseball player catches a fly ball, it is argued, they don't compute the trajectory of the ball's flight and run to correct spot. A Player is computing trajectory of ball to grt there. Rather, they keep moving so that they are in the right relationship with the ball. Like you could just say something like, the player is continuously aligning their body with the ball in such a way that as the ball falls, they're relating to the ball in the right way physically, such that eventually they are where the ball, they need to be to catch the ball. That second description, Epstein argues, is not a computational description, it's not a description of anyone computing anything, rather it's a description of how a body exists in relation to other bodies in the world, as this idea of embodied cognition, rather than merely computational cognition, This, it is argued, is not computational but interactive.

Describe different plausible objections to the computational theory of mind.

Epstein's paper about embodied theories of mind provided objections to computational theory of mind. This was this idea that to be cognitive, to think, it's not enough just to compute. We have to be in the world. We have to have bodies. So this example of the kind of thinking involved when a baseball player sees a fly ball and is able to run to the right spot and catch it. The idea is this sort of process, it's a process of thinking. It's a process of cognition, and yet it's not purely computational because it requires the body in the world to make it happen. And the Turing machine, it's argued, could never perform these computations, could never actually implement the process of running down that fly ball because it's essentially embodied and a Turing machine doesn't have a body. You can come up with all kinds of candidate cases of similar examples where we might think that to be cognitive is to be in the world, to be embodied and in terms of Turing machines , they fundamentally fails to grasp this important aspect of cognition. Computational theory of mind, very simply conclude to the idea of the computer is a Turing machine The theory puts forth hat computers may be more complicated, they may implement things in all sorts of ways, using electrical signals, maybe if our brains are computers, they're using neurological processes, but everything it does, you might argue, could ultimately be represented as a Turing machine. Additional arguments to the computational theory of mind are Argument from creativity. There are all these things that we know we do where we don't have good Turing machine implementations, things like composing music, writing novels, et cetera, it's unclear how these can all be understood in terms of Turing computable functions. Argument from abduction A big blind spot in the Turing approach to cognition that is missed because it doesn't account for inference to the best explanation. More so than actually computing functions, figuring out what explanations are good explanations where the goodness of an explanation is almost an aesthetic appreciation for the quality of that explanation rather than any sort of precise, mathematical property of that explanation Might lead one to think that it

Describe the methodological issues faced by fMRI studies, especially the issue of "hypothesis-free science," and the idea that "correlation is not causation."

FMRI are expensive to run so they can not run large sample sizes Often we can take a lot of measurements from one person, but recruiting a lot of different people, a lot of different patients or participants can be quite expensive. A conceptual issue is that fMRI studies often are accused of engaging in what's called hypothesis -free science, meaning where a typical way of doing fMRI research, has been to observe cognitive phenomena and then go out and find a neurological cause or explanation often without any hypothesis to guide where that neurobiological explanation will be. In most of science, you have a hypothesis, an idea based on previous studies, based on observation, even just based on your hunchesl you go out and you test that." But in much fMRI work it's often the case of, "Well, we're just interested in what parts of the brain are lighting up when this cognitive activity occurs." So there's no hypothesis there. There isn't a point where you say, "Okay, it's going to be that part of the brain, because we have so little theory to guide us." Quite often we just want to make the observations. It's often alleged FMRI doesn't really tell us how cognitive algorithms are implemented, just where they are implemented There's a departure from the traditional scientific method of hypothesis testing that should at least give us some methodological pause about how much we can learn from a kind of study where we're just sort of observing what happens rather than actually first coming up with a hypothesis and then testing that hypothesis. A second big issue is that correlation is not causation. People on a sidewalk using their umbrellas and cars in that same street using their windshield wipers are NOT correlate, and there's no causal connection between the two, because if all the people on the street suddenly put down their umbrellas, it wouldn't necessarily make the cars stop using their windshield wipers. The people on the street would just get wet. Similarly, if the cars all stop using their windshield wipers, people on the sidewalk certainly wouldn't be inclined to just put down their umbrellas, but rather, the cars using their windshield wipers would just no longer be able to see the road. The reason is that

Describe the main idea/argument of the Green, reading, the main reasons for reaching their conclusion, what findings or ideas they describe, and what the significance of those findings or ideas to be

Greene et al. discusses emotion and morality. They argue two versions of the trolley problem (avoiding killing people, kill one or kill 5); one involving flipping a switch, the other involving pushing a person from a bridge. fMRI studies found that when participants considered what to do in the person-pushing case, the parts of their brains associated with emotional processing were especially active. This did not occur as often when the considered the case that involved flipping a switch. This may explain why some people see the two cases as morally distinct. Regions of brain recruited in solving second problem associated with emotional response.

Describe the main idea/argument of the Jonas and Kording, reading, the main reasons for reaching their conclusion, what findings or ideas they describe, and what the significance of those findings or ideas to be

Jonas and Kording say that it may be an important intermediate step for neuroscience to develop methods that allow understanding a processor, right, where a processor here is just, you know, as a computer, a circuit board processor for computing some function. They say, you know, having methods that allow us to understand how such a processor actually computes rather than just where it computes might be an important intermediate step for neuroscientific understanding. If we study just the flow of electricity in a micro-processor like the kind that is used to to implement classic video games, we are not able to get a high-level understanding of the game that is being implemented. Even if we know everything about exchange of electricity in a nintendo cartridge we don't know the game, what is seen on the screen. By the same token/reasoning, careful study of the flow of electricity in the brain (essentially, what fMRI gives us) may not be sufficient to full understand what the brain is implementing at a higher level. Jonas and Kording say that "it may be an important intermediate step for neuroscience to develop methods that allow understanding a processor?" There are myriad of ways in which a microprocessor and the brain are similar including hierarchical organization with different functional modules, the ability to exhibit localized functions, and the interaction of multiple components involved. The key criticisms of the IP metaphor explain why the in-depth understanding of a microprocessor cannot provide us with the same level of understanding of the brain processes.

Describe the main idea/argument of the Lande, reading, the main reasons for reaching their conclusion, what findings or ideas they describe, and what the significance of those findings or ideas to be

Lande argues the idea that the brain is a computer is a useful metaphor because it enables us to develop hypotheses, often mathematically precise hypotheses, about what the brain does that can be tested, refined, and re-tested. Because of this usefulness, he concludes, the metaphor of the mind as a computer is effectively "true." , Lande uses some examples from the history of physics to support his conclusion. In the Church-Turing discussion, the author notes how scientists like Turing discovered ways to make computers do "computations" in a variety of ways and with increasing efficiency. This was seemingly used as evidence that the human mind and computers are comparable in certain (admittedly imprecise and metaphorical) ways. However, this seems to fall into the trap that Epstein discusses in his essay "The Empty Brain." While we can begin with two correct premises (computers can calculate and process and behave intelligently), it is a stretch to then conclude that we must behave in a very similar way because we too can process information

Describe the main idea/argument of the Marcus et al, reading, the main reasons for reaching their conclusion, what findings or ideas they describe, and what the significance of those findings or ideas to be

Marcus et al argues that the brain must be studied in an inter-disciplinary way, guided not just by neurobiological observations but also by theories. We need to bring psychology, artificial intelligence, and philosophy to bear on the understanding of the brain. Theory takes various forms studied under various subjects, e.g. in computer science, theory means construction of algorithms Neuroscience progress and analysis are to be vague, slow, and antediluvian. Marcuses mentions of how neuroscience is highly experimental and not theoretical are significant to general interdisciplinary study of cog sci " Theory is practically unstudied, and large scientific funds for research go to experimentalists because building theories is considered risky and requires interdisciplinary training but this can be resolved by giving to the NSF and NIH to fund centers geared to help foster an academic environment for researchers from several disciplines. neuroscience progress and analysis to be vague, slow, and antediluvian (9).

Describe Nagel's What Is it Like to Be a Bat? intending/purpoting to establish an "explanatory gap" between physical descriptions of the mind and phenomenal consciousness and a counter argument to Nagels thought experiment

Nagel argues that the genuine knowledge of the subjective experience of a bat is fundamentally inaccessible to us.There are facts about the conscious experience of bats that are not communicated by any physical description. We could know all of the physics of what it's like to be a bat, all of the biology, and there would still be a kind of a missing fact there, namely the fact about what it is like to actually be a bat. Conclude that there are facts aboutconscious experience that are different in characterto facts about physical phenomena, and some of those facts (e.g., what it's like to be a bat) may be inaccessible to humans. So as I've just said, even the most intricate, detailed description of bat neurobiology, it is argued, would not reveal anything about the subjective experience of being a bat. Relating to the explanatory gap., even if we know everything physical, there is to know about the operation of the bat's mind, how the bat cognitively gets around in the world,there is a further fact that is not explained by that physical description,namely the fact of what it's actually like to be this bat, what that bat's qualitative experience is like. So some of the facts about what it's like to be a bat, some of the facts about consciousness might be inaccessible to humans, right? The idea of the explanatory gap is even full understanding of our own consciousness, why we have the experience we do, why our experiences have the phenomenal qualities that they do, may in some sense be inaccessible to us as well. And so to the extent that consciousness is a part of the mind, you might think the conclusion goes, well, then the mind is not necessarily entirely physical. Smart, makes a counter argument to Nagels thought experiment as he ultimately thinks that when we report experience of sensations, we're reporting experiences of brain processes. Churchland, also counterargues that we should reject Nagel's inference from the inconceivability of a physical explanation of bat-experience to the existence of facts about bat experience that are not physically explainable. She would argue that we should reject Nagel's inference from the inconceivability of a physical explanation of bat experience to the existence of f

Describe how optical illusions illustrate the cognitive impenetrability of mental modules.

Optical illusions illustrates the cognitive impenitrability of mental modules by using visual perception (eg. visually percieving ML lines as different lengths) and propositional knowledge (eg. you know the proposition that the ML lines are the same length) which is meant to demonstrate they are separate modules and understanding those parts of optical illusions to exemplify the idea that the visual module here is being shown to be cognitively impenetrable/informationally encapsulated where taking in new information from ones propositional knowledge module cannot override the functioning of their visual system, which is picking up on (for ML) cues given to them by the length, the directions of these arrowheads,etc that tell the visual system that these two red line segments are different lengths. An optical illusion exemplifies the idea that a visual system is never going to convince me to revise my propositional knowledge. We can look at cognitive impenatribility by propositional knowledge as all the propositional knowledge in the world can't make you not fall for the illusion, showing the idea of your visual system as cognitive impenitrabile by the other (cognitive module). We can use the Muller-Lyer (ML) illusions to help desribe how optical illusions in general illustrate cognitive impenetrability of mental modules. We can see the bottom line as longer than the top one (with >---<. Vs <—>)

Describe in depth how we use pixel-clustering to see distinct objects in our visual field.

Pixel cluttering is this idea of how we see pixels in our visual field that are similar to one another in color as produced by the same object/foreground/background, whereas two differently colored pixels are treated as distinct objects and we clusters the similar colors we see together and the distinct colors we see separately in order to perceive different objects or different patterns within our visual field. Relating to the computational level of Marr's levels of analysis you can think of one thing you are doing computationally when you are looking at a chair with a, say green cushion, on brown wood is identifying clusters of similarly colored pixels and other clusters of differently colored pixels. Light reflected on one part of an object, say a basketballs lines being darker, than the rest of the ball, we can cluster it visually The more subtle the differents, the harder it is to see distinct objects in our visual field. this was a key innovation in thinking about vision.

Describe the main idea/argument of the Robbins, reading, the main reasons for reaching their conclusion, what findings or ideas they describe, and what the significance of those findings or ideas to be

Robbins distinguishes between "input system modularity" (what does it mean for mind tome modular- different systems treat different inputs- visual goes to one place, auditory goes to another) and "central system modularity" or "massive modularity" (everything the mind does is modular-math module, verbal module, etc. ). He Provides reasons for being skeptical of both. Against input system modularity, Robbins points out how one cognitive module sometimes "fills in the blanks" for another. There is not an even split in the mind between different parts that treat different inputs. Against massive modularity, Robbins argues that in many people, their cognitive abilities are correlated. People who are good at verbal reasoning tend to also be good at mathematics and to also be good at spatial reasoning. O One challenge to modularity robbins gave was a description of how cognitive neuroscience studies have disputed the concept of input system modularity. Because It seemed to me that the idea of anatomical localization of neural processing is central to his argument. Robbins explained how fMRI studies of systems such as speech production and face processing have demonstrated that brain regions previously thought to be specialized for particular functions are actually involved in many processing systems ("massive redeployment hypothesis"), contradicting perhaps input system mosularity. Also described a study that showed that the average cortical region is actually implicated in the processing of nine domains!

Describe the main idea/argument of the Smart, reading, the main reasons for reaching their conclusion, what findings or ideas they describe, and what the significance of those findings or ideas to be

Smart argues that when you report a conscious sensation (e.g., "I am seeing a red curtain right now"), what you are actually doing is reporting a process that is occurring in your brain. You are describing, in a shorthand and somewhat confused way, the process by which red light is hitting your eyes and being interpreted by your brain as a red curtain. He reaches this conclusion because he believes it is the simplest explanation of what is happening when we report our conscious sensations. Anything else when we are reporting seeing a red curtain would have to suppose that there are additional things happening besides what is happening inside our brains that explain conscious experience. Occam's razor is the idea that simpler models all else being equal or better means that you should err on the side of just thinking what what I'm saying what I'm talking about when I talk about my conscious perceptions. Conscious sensations is really are just brain processes

Describe all three of Marr's levels of analysis for an information-processing system. Give examples of how the levels can be used to describe an information-processing system.

THERE IS THE Computational Level: What needs to be done and what function must be computed to accomplish that goal? Functions, a plan Algorithmic Level: How should inputs to the function being computed be represented, and what algorithm should be used to compute outputs from inputs?.Procedures and Steps An algorithm is is an abstract description of a set of steps for solving a problem, for computing a function. Implementation Level: By what physical process are the steps specified in the algorithm carried out? (outcome baby) FOR A CASH REGISTEER EXAMPLE Marr investigated how can we actually get a computer to see the way that human beings do? To explain, he uses the cash register example. Computational Level: The cash registers arithmetical function of addition. We want to do use the properties of addition to help us get a finalized bill. We can look at the zero property, buy something and nothing is the same as buying something, as one example of these properties. Algorithmic Level:What procedure does the cash register use to compute the addition function? What algorithm the cash register uses to calculate outputs (sums), from inputs (pairs of numbers), Representing machine functions, for example, in roman numerals What an algorithm ultimately is, is an abstract description of a set of steps for solving a problem, for computing a function. Implementation Level: This is doing an algorithm for computing something. How does the register actually implement an algorithm for computing the addition function? What process are algorithmic steps actually plate out? The answer here depends on the kind of cash register that we're talking about: For example, A human could act as a cash register, using a pen and paper to implement the addition algorithm, or (if they were really good at mental math), using the the neurons in their brain. Within the domain of neuroscience this is described as the mechanical lebel looking at what are neurons in brain doing to take steps. Once you've decided that any info processing system can be understood at either a computational, algorithmic, or implementation level, you really can't address all three levels while staying within one traditional discipline hence cogsci is interdisciplinary.

Describe in depth the structure-from-motion theorem and how it facilitates three-dimensional vision.

The Structure in motion theorem, by Ullman (a student of Marr) is that for any four points on a moving rigid body that are not on the same 2-D plane (differ in respect to dimensions), we can uniquely determine the 3-D structure of those points (distances of straight lines between the points) and the direction of movement of the body from three unique orthographic dimensions. So as if we were to take any three-dimensional object, and chunk it up into "rigid" bodies, trictly speaking, the same straight line distance, the same angles should maintain between all the different points within a rigid body. It is rigid, it can't sort of warp and shift as it moves. If such a body moves, all its points must maintain the same straight-line distance from each other. This facilitates 3D vision because it basically puts forth the idea, like that if there's a three -dimensional body moving in space I can see it in two dimensions, but if seen by three different angles, from three different perspectives, then those three perspectives tell me uniquely what the full three -dimensional structure of the body has to be, even though I'm only seeing it in two dimensions. If we project something it onto our eye ad our eye's two -dimensional visual field in three different ways. We'll know what we are looking at is a 3D structure. Rigidity assumption is key to this ability, and keeping this is a rigid body. Our visual field takes something that lives in three dimensions and turns it into a two -dimensional projection. chunking these objects up into sets of four coordinates that we decide those things are always gonna move together. Those things are always stuck together Related to Marr's algorithmic level of computation because if a computer could chunk things up into rigid bodies and in further 3D structures so evidently can our brains and perhaps so could other things. Perhaps you could with pen and paper, get these three coordinates, and then figure out what 2D structure would have to have to have generated them, right? That would also be an algorithmic implementation of that same thing, but the fact is, that's the algorithm that we know that we have to run. We have to find an algorithm that achieves that computational goal. This i

Describe how an fMRI takes advantage of the role of blood in the functioning of the brain to provide accurate images of brain activity (hint: more, less, slower, longer, stronger).

The brain is full of blood vessels, there is blood pulsing through your brain constantly, and when certain neurons fire, more oxygenated blood cells travel to that neuron. When a neuron fires, it recruits oxygenated blood towards it, and that blood flows towards that neuron. Oxygenated blood cells make the molecules in a brain region less susceptible to the magnetic field and because there's more oxygen, the effect of the magnetic field is less strong in that region. And this means that when molecules in an oxygenated region of the brain are perturbed by radio waves, they're slower to return to a non -perturbed state,, so when there's a region of your brain that has a lot of blood flowing towards it, because those neurons are firing, inside the cell body, you've got electrical pulses that are being sent along the axons and received at the dendrites of other neurons. What's being received at the receptors of the other neurons is either just raw electricity or in some cases, molecules that are being sent that are being activated by that electrical signal and sent to other neurons. And all of this is just sort of creating electrical activity in certain parts of the brain.It's recruiting that oxygenated blood to those parts of the brain and is detectable within an fMRI machine. So when we talk about neurons firing, we're talking about this process of message passing, molecular message passing between the neurons in the brain. Inside a magnetic field, hit with radio waves, it'll change the spin of the protons within that region of the brain, but then those protons will be slower to come back to their non -perturbed state from when they were hit with the radio waves to when they weren't. The parts where they're going back to the non-pertrubed stateslower, that's where the oxygenated blood is, and that's where the neurons are firing, An FMRI takes advantage of the role of blood in the functioning of the brain to provde accurate images of brain ativity as when a neuron fires because as... More oxygenated blood travels to that neuron And Oxygenated blood cells make the molecules in the brain less susceptible to the magnetic field When the molecules in an oxygenated region of the brain are perturbed by radio waves they'r

Describe in depth the four main neuroscientific attempts to define a physical implementation of consciousness.

The four main neuroscientifc attempts to define a physical implementation of consciousness as described by Seth and Bayne identify the following neuroscientific theories to define a physical implementation of consciousness and attempt to explain the feeling that we all have that there is something that it is like to have our experience in virtue of neural phenomena, brain phenomena,: Higher order theories (HOT) The higher order theories focus on the ideas that brains achieve consciousness (the result of the brain meta -representing first order states) when and because they are capable of higher order representation. So if we want to ask is an organism conscious, we can now reduce that to the neuroscientific question of is this organism capable of metarepresentation, of higher order representation. If you have some representation of your environment, a map of your world, or a representation of how that tree outside grew out of the ground over 100 years,then you have a representation of that representation, also stored, implemented, elsewhere within your brain. So the brain kind of represents its own processes. It thinks about itself. A lower representation might be my image of the lecture hall in front of me, right here A higher order representation is my representation of the fact that I have a lower order representation of the lecture hall in front of me, right? I can think not only about the lecture hall in front of me, but the fact that I can thinking about the lecture hall in front of me. And the thought is that second order thinking about thinking might be perfectly instantiated within the brain, right? Global workspace theories(GWT) Global workspace theories say that consciousness is instantiated in a hub that is accessed by a host of other faculties and that the brain also contains structures, neural architecture that acts as a hub where the different faculties (relating to moduarity) are coordinated. Taking all the tools in our cognitive toolbox and using it simultaneously. The ability to sort of perceive your own thinking is necessary to have a global workspace and coordinate between these different activities all at once. While I'm simultaneously gesturing with my hands in a certain way, I'm doing a

What is the significance of Turing machines for cognitive science?

The turing machine is very signigicnt to cognitive science, especially to cognitive science argument that mind is computer. The idea of the mind as a computer can be understood as the mind as a Turing machine; The significance of this to cognitive science is it supports the idea that if an algorithm can be implemented on a Turing machine and a mind is a computer,then we should be able to describe anything that the mind does as a Turing machine. So one conclusion you could draw from this is, well, the mind simply can't say of any theorem, of any sort of mathematical sentence, whether it's a theorem or not, whether it's a tautology or not. He proposes the first really thought out computational theory of mind from his work with the entschediums problem Turing worked on Entschedidungs problem looking can we identify any tautology of first order logic; his answer was no There are sentences that have that form such that no Turing machine can ever tell you whether they're true or not.It cannot tell you whether, given some description, perfectly well -formed description of another Turing machine, whether that Turing machine will loop forever or halt. And Turing just proved this, like it's just a logical fact. So his idea is to come up with a very general architecture for a machine that could compute any function Except for those for which no finite algorithm is capable of computing every output for an input We can say of some Turing machines whether they halt or not, but there will always be Turing machines that we can't say whether they halt or not, and so a first order logic sentence that's about whether that Turing machine halts, we wont know whether it's a tautology or not. So his idea is can I define a machine that is so general that that machine can do Anything that any machine that could compute any function can do ; that it could compute any function that could possibly be computed and Obviously would then anything that it could not compute would therefore be sort of strictly uncomputable The goal that he sets himself is to define this very general architecture and he is able to do so.

Describe Jackson's "Mary's Room" thought experiment. intending/purpoting to establish an "explanatory gap" between physical descriptions of the mind and phenomenal consciousness and describe a counter argument to Jackson's thought experiment

Thought experiment meant to establishidea of consciousness as not fully physical Mary's room experiment was where Mary is sat in a black and white room. She's lived in this black and white room her entire life. But in this black and white room, she has been given the finest education in the world, She has chosen to become a scientist of color, in the sense that she is going to understand color vision and color perception better than anyone ever has. She learns about all of the spectrumspectrum of light that the human eye can see. She knows everything there is to know about the human eye. She knows everything about optics. She knows everything about how the eye connects the brain. She knows everything about the conditions under which different human beings report seeing a green object, report seeing a blue object, report seeing a red object. There is nothing in the physical description of that process process that is not known to her. She understands color blindness. She understands why a shade of blue might look darker to me than it does to you. She knows it all, DESPITE THISt, there comes a moment where she is taken out of this room in which everything is black and white and she sees, for the first time, a red object the experienc of seeing red was different than expected. Connects to Idea of the knowledge argument where when you experience something you gain new knowledge of what its like to have that experience. (Description of horror movie wont be the same as experience of seeing movie. From having that experience you're going to have new learnings: experience, new emotions, new ways of sort of relating objects to one another. Explanatory gap in the real experience vs the idea/thought of/planned experience A counterarguent to Jacksons thought experiment is there isn't actually something new learned in the experiment as its just the same knowledge in a different mode of presentation, which is not the same as learning something new because what's at stake is the questiono f whether there are facts about phenomenal experience and facts with things that can be learned that can't be captured by just the physical understanding of light and color and all those things that go into color vision as well.

What do neural networks have to do with connectionism?

What neural networks have to do with connectionism is with neural network connections, you've got weights, you've got neurons connected to other neurons, you've got weights to the connections and all of that together determines the values taken by all the neurons. Connectionism aims to explain cognition with these neural network. Chat GPT is able to use a sophisticated neural network architecture (key word in connectionism), but a neural network architecture. Connectionism tries to exemplify the power of neural networks to master cognitive tasks.

Describe the advantages of fMRI over earlier brain-imaging methods, and the drawbacks of current fMRI technology.

vfMRI provides a method for observing brain activity . Looking at brain in vivo is best way to mind read. Advantages: It is aafe and non-invasive, even for kids., Unlike pneumoencephalography (PEG) of 1920s that was extremely painful. It drained via lumbar puncture and replaced cerebtosphnal fluid with gas like air. Safe for repeated and extended use, can observe a patient doing the same cognitive process multiple times and track brain activity. You know, you can go in an fMRI for a long period of time, you can do it again and again. We can observe a patient doing the same cognitive process multiple times and track brain activity. So we can get repeated samples, Unlike CT that is claustrophobic for patients; limiting extended use No special preparation (just can't have metal-use CT if metal). Simultaneous view of the brain and its function. By tracking blood flow within the brain, which is essentially what fMRI is doing, we're able to see which parts of the brain are functioning most strongly while also just getting really nice images of the brain, like we can see these wonderful cross sections of the brain where you can really start to see its finer structure in far more detail, and it has a high degree of spatial resolution. High degree of spatial resolution. With MRI, we were able to see the brain for the first time and what might be wrong with the brain but we werent seeing function because we weren't being shown an image of, what the part of the brain that's sort of working or most active That's where the functional in functional MRI comes from. Disadvantages: The fMri has poor temporal resolution, as we basically have about one second time frames. The FMRI works because of a slower period of procession in the oxygenated parts of the brain, leading to longer emissions of waves from parts of the brain that are oxygenated at a given moment,, itworks off of there being a certain length of time scale that we have to wait in order to get the signal that's enabling us to sort of figure out what parts of the brain are firing when. This is a problem because, with fMRI, we have a temporal resolution of about one second time steps. We can say,, here's what this person was doing at this second, here's what this per