This post will go over Dennett's views on animal experience (and to some extent, animal intelligence). This is not going to be an in-depth exploration of Daniel Dennett's thoughts over the decades, and will really only focus on those parts of his ideas that matter for the topic. This post has been written because I think a lot of Dennett's thoughts and theories are not talked about enough in the area of animal consciousness (or even just intelligence), and more than this its just fun and interesting to me.
It is worth noting that Dennett is known for being very confusing and esoteric and often seeming to contradict himself. I had to read not just his own writings but writings of those who knew him or attempted to ironman his ideas to put this together in this simplified form. Which is essentially what I'm also doing right now. There is not a lot of 'essay structure' so it ends quite abruptly and may be a little disjointed but I've tried to make it as comprehensible as possible.
I am not writing this to directly critique or support his views, although I may make a comment on a claim of his when I feel like it (or whatever, really).
I would really appreciate kind feedback if you felt it was too ramble-y, or used too much X, or should've done X, or so on, instead of silent downvotes.
Drafts and Fame
The brain is highly parallel, with many streams of input being processed simultaneously and many aspects of input processed simultaneously (such as colour and movement being processed in separate streams in the visual cortex). Different parts of brains track different inputs, different properties (i.e. V4 for colour, V5 for movement). Some of these signals survive (are 'popular') but many are unpopular and die out. This is what Dennett calls 'fame in the brain'.
Through this process, the brain makes an draft (interpretation / predictive model) of its inputs (Dennett uses different language, but it boils down to the same thing, as I usually opt to use the general common use terms instead of his own specialised vocabulary). Popular drafts go on to influence other modules, making their impact on memory, behaviour and cognition and just generally sticking around, unpopular ones do not.
Drafts can always be revised as new interpretations of input are created and selected. For example in the colour phi illusion (an illusion where a red and blue dot flashing slightly delayed across a screen creates an illusion of motion), the initial draft would be simply a blue dot, then perhaps a 'blank screen and red dot' draft would be created, until eventually (a few hundred milliseconds after it begins, probably) a 'blue dot moves from top to bottom changing colour' draft gets created before quickly winning in popularity, becoming globally distributed (see the foot notes for reference to global workspace theory) and dominating memory, report, so on.[1]
Autostimulation
To summarise Dennett's view of language. Proto-humans developed more and more complex communication, learning a proto-language, and crucially they gained the ability to ask questions. At some point, they accidentally started asking themselves questions. In the process of asking themselves a question, the answer to that question, which was hidden somewhere in a part of that proto-humans brain inaccessible to use, suddenly became accessible to use as the question made its way into their ear, stimulating their brain, drawing out an answer. This self-asking is called 'probing'.
This was the beginning of autostimulation, of which asking yourself questions ('self-interrogation') is the strongest and most blatant, as it comes up in the lives of pretty much all humans all the time. From this came talking to yourself, self-narration. The ability to speak to yourself in this way was highly adaptive, so the sorts of connections involved in autostimulation were genetically encouraged, the basis for both self-and-other communication laid down in the brain before birth.
As brains learned to talk to themselves, this became an ingrained neural habit, persisting even in the absence of explicit linguistic or reflective thought (even those with global aphasia maintain a very real capacity for autostimulation even if it is impaired). Because of this, the structure of language (such as lexicon and syntax) influences the brain even subconsciously.
The Joycean Machine
Dennett calls the language faculty privileged / special. This is because whatever you report to yourself (report out loud or silently) will necessarily be heavily selected for and globally broadcast and especially importantly rehearsed (when you report to yourself, that information becomes very long-lasting and stable), even if the signal wasn't globally broadcast before it became reported. Essentially the language faculty allows some signals a free-win, and so whatever draft of a thing is accessible to the language faculty will be the draft that people feel was the true experience, as when prompted on what they experienced as it will be the one they report, and thus the one they remember and continue to think about.
When someone's streams of input is 'probed' (they ask themselves or are asked what there experiences were), a single most-popular draft becomes 'precipitated', selected for above all others, canonized, in a sense. This provides a 'time line; a subjective sequence of events from the view of an observer'. This time line allows us as scientists to compares peoples subjective timeline to other timelines such as the timeline of neurological processing which may or may be match up with the subjective one.
A very core idea of Dennett's is that there is no facts about experiences independent of probes, since without probing there are no privileged drafts in the brain, only many different ones. Probing is important because as mentioned it involves a kind of sustained rehearsal and global distribution of contents that language facilitates.
However autostimulation exists even without a functioning language faculty, and is constantly everpresent in the brain, meaning some contents are still being 'rehearsed', providing a sort of sublinguistic time line. Essentially in the brains of humans who lack the language faculty but retain autostimulation, they are still sub-linguistically probing themselves (or rather, continuing to act and think under the influence of probes on neural habit: with some contents rehearsed and globally distributed as if they were in language).
Dennett's original view was that conscious states were nothing more than popular states, since then he's seemed to change his mind a little referencing meta-representations as essential for consciousness[2]. Regardless, he basically claims that whatever contents are being globally rehearsed and processed are the contents of your consciousness. He calls the contents of consciousness the contents of the Joycean machine (which he calls a serial virtual machine loaded onto parallel hardware), as it contains only those contents which are being rehearsed in a special, global way facilitated by language.
Most of the content of the Joycean machine is non-linguistic, but it is the influence of language that facilitates the existence of the machine.
Animal (un)consciousness
Dennett gives two core reasons why he sees human consciousness as so profoundly different from animal consciousness (+ mostly from young human baby consciousness too), and thus why he doesn't see animal suffering as very serious.
Selfless
The first is that non-human animal minds lack a self model. They have no concept of 'I' woven in their brain, and because of this they have no narrative centre of gravity, no meaningful model of themselves as an individual in the world. They are perpetually disassociated. Dennett makes reference to disassociated traumatised children who report no longer or mostly no longer feeling pain due to their disassociation, using this as an indication that animals might be the same way as they never even develop a self to dissociate from in the first place. [3]
He then says in another article, for 'body protecting states' to matter, there must be an 'enduring, complex subject for whom they matter because they are a source of suffering'. In thinking about it this way, bodily damage induces a whole extra set of judgements / interpretations unique to humans, judgements that in Dennett's view constitute suffering. Put another way, the experience of pain to a creature with no narrative self is so fundamentally different in his view it isn't really comparable to humans, where mostly everything we recognise as suffering exists only because of the narrative self.[4]
Unified mind
Daniel Dennett consistently claims that non-human animal minds lack unity, as although they processes information, without the 'timelines' provided by autostimulation, there is never any point where the organism as a whole comes to perceive anything, there are only parts of organism's brains working with different information.
He mentioned his precise beliefs and predictions here, saying that long distance reverberant / re-entrant interactions have to be laid down by language use, and that without these connections there is no functional unity to the nervous system.[5]As a result, strictly speaking, there can be nothing 'it is like' to be a cow, for example, since there is no cow-system, only different sub-systems within the cow with their own shifting loci of activity.
He also seems to argue against the functional unity of representations in non-humans minds as well. In one example he points out that if you train a rabbit to be fearful of something that its seeing with its left eye, this won't generalise so when the rabbit sees the cue on its right it won't be scared. 'What's it like to be a rabbit? Which one, the left or the right?' he says, implying their disjointed reactions underlie a disjointed awareness of the world. [6]
In another example he points out research showing that snakes use 3 separate modalities to catch mice, always using the visual system to strike and locates it via smell, and then devours it via tactile information. Notably, according to the article even if the mouse dies right in front of the snake it will still follow the smell trail anyway. They also point out that this behaviour is especially some snakes like boas and pythons, where even though the snake 'must have ample proprioceptor-y information' about the location of its prey, it searches for it in its grasp with only its scent anyway. From this he claims the snake lacks a central representation of the mouse its chasing.
Presumably then language learning in his view doesn't just allow for a unified conscious mind, but also allows reliable central representations of objects in one's world, mind and body (it facilitates general schema building).
Pretty much all this information comes pretty directly from his book Consciousness Explained, but you can also find it mentioned it most of his writings.
It seems like an additional power of language for Dennett is the power of meta-representation, as language allows you to work with your own representations.
If you're interested in this it might be worth reading this report, where Kaj seems to report a pretty significant dissolution of the narrative self while retaining otherwise normal functioning, finding a significant drop in negative affect and most of all aversiveness of negative experiences. There was a dissolution of craving and a breakdown of a 'subjective glue' binding things together. I don't really know what this means but do wonder if this could imply, contrary to the Buddhists, that animals are actually more protected from cravings than humans are and thus do suffer less as they become less entangled in things.
In regards to global workspace theory, my best guess is that he believes the all-or-nothing behaviour associated with the global workspace (where a consensus is reached within the workspace on interpretations of input via reverberatory / re-entrant activity) is caused by the influence of language, whereas non-human animals only have contents that are in varying stages of globally distributed without any of these being privileged. This tracks with his strong association between autostimulation and re-entrant connections and the very high importance of long distance re-entrant connections to the functioning of the global workspace.
I found this example very unconvincing, as its not like the rabbit understands why its been conditioned into being fearful in the first place, so it easily could have a unified mind but simply feel a involuntary jump of fright at seeing it one eye but not the other.
Pretty much all this information comes pretty directly from his book Consciousness Explained, but you can also find it mentioned it most of his writings.
It seems like an additional power of language for Dennett is the power of meta-representation, as language allows you to work with your own representations.
If you're interested in this it might be worth reading this report, where Kaj seems to report a pretty significant dissolution of the narrative self while retaining otherwise normal functioning, finding a significant drop in negative affect and most of all aversiveness of negative experiences. There was a dissolution of craving and a breakdown of a 'subjective glue' binding things together. I don't really know what this means but do wonder if this could imply, contrary to the Buddhists, that animals are actually more protected from cravings than humans are and thus do suffer less as they become less entangled in things.
In regards to global workspace theory, my best guess is that he believes the all-or-nothing behaviour associated with the global workspace (where a consensus is reached within the workspace on interpretations of input via reverberatory / re-entrant activity) is caused by the influence of language, whereas non-human animals only have contents that are in varying stages of globally distributed without any of these being privileged. This tracks with his strong association between autostimulation and re-entrant connections and the very high importance of long distance re-entrant connections to the functioning of the global workspace.
I found this example very unconvincing, as its not like the rabbit understands why its been conditioned into being fearful in the first place, so it easily could have a unified mind but simply feel a involuntary jump of fright at seeing it one eye but not the other.
This post will go over Dennett's views on animal experience (and to some extent, animal intelligence). This is not going to be an in-depth exploration of Daniel Dennett's thoughts over the decades, and will really only focus on those parts of his ideas that matter for the topic. This post has been written because I think a lot of Dennett's thoughts and theories are not talked about enough in the area of animal consciousness (or even just intelligence), and more than this its just fun and interesting to me.
It is worth noting that Dennett is known for being very confusing and esoteric and often seeming to contradict himself. I had to read not just his own writings but writings of those who knew him or attempted to ironman his ideas to put this together in this simplified form. Which is essentially what I'm also doing right now. There is not a lot of 'essay structure' so it ends quite abruptly and may be a little disjointed but I've tried to make it as comprehensible as possible.
I am not writing this to directly critique or support his views, although I may make a comment on a claim of his when I feel like it (or whatever, really).
I would really appreciate kind feedback if you felt it was too ramble-y, or used too much X, or should've done X, or so on, instead of silent downvotes.
Drafts and Fame
The brain is highly parallel, with many streams of input being processed simultaneously and many aspects of input processed simultaneously (such as colour and movement being processed in separate streams in the visual cortex). Different parts of brains track different inputs, different properties (i.e. V4 for colour, V5 for movement). Some of these signals survive (are 'popular') but many are unpopular and die out. This is what Dennett calls 'fame in the brain'.
Through this process, the brain makes an draft (interpretation / predictive model) of its inputs (Dennett uses different language, but it boils down to the same thing, as I usually opt to use the general common use terms instead of his own specialised vocabulary). Popular drafts go on to influence other modules, making their impact on memory, behaviour and cognition and just generally sticking around, unpopular ones do not.
Drafts can always be revised as new interpretations of input are created and selected. For example in the colour phi illusion (an illusion where a red and blue dot flashing slightly delayed across a screen creates an illusion of motion), the initial draft would be simply a blue dot, then perhaps a 'blank screen and red dot' draft would be created, until eventually (a few hundred milliseconds after it begins, probably) a 'blue dot moves from top to bottom changing colour' draft gets created before quickly winning in popularity, becoming globally distributed (see the foot notes for reference to global workspace theory) and dominating memory, report, so on.[1]
Autostimulation
To summarise Dennett's view of language. Proto-humans developed more and more complex communication, learning a proto-language, and crucially they gained the ability to ask questions. At some point, they accidentally started asking themselves questions. In the process of asking themselves a question, the answer to that question, which was hidden somewhere in a part of that proto-humans brain inaccessible to use, suddenly became accessible to use as the question made its way into their ear, stimulating their brain, drawing out an answer. This self-asking is called 'probing'.
This was the beginning of autostimulation, of which asking yourself questions ('self-interrogation') is the strongest and most blatant, as it comes up in the lives of pretty much all humans all the time. From this came talking to yourself, self-narration. The ability to speak to yourself in this way was highly adaptive, so the sorts of connections involved in autostimulation were genetically encouraged, the basis for both self-and-other communication laid down in the brain before birth.
Due to its adaptiveness, the pathway was shortened from brain -> mouth -> air -> ear -> brain, to brain -> brain, so now people could auto-stimulate without having to reveal anything out loud. From this was born inner-speech, internal explicit autostimulation. And partially perhaps from inner speech came things like mental imagery, which Dennett says may have been derived from things like drawing diagrams. Due to autostimulation's ability to boost the popularity of relevant information in a subsystem, it acts almost like a control system, allowing for attention to be focused and behaviour controlled. Autostimulation also allowed for / greatly facilitated rehearsal, autobiographical memory, creativity, reason, mental rehearsal of action and more, including developing a sense of yourself over time and a complex theory of mind, according to Dennett.
As brains learned to talk to themselves, this became an ingrained neural habit, persisting even in the absence of explicit linguistic or reflective thought (even those with global aphasia maintain a very real capacity for autostimulation even if it is impaired). Because of this, the structure of language (such as lexicon and syntax) influences the brain even subconsciously.
The Joycean Machine
Dennett calls the language faculty privileged / special. This is because whatever you report to yourself (report out loud or silently) will necessarily be heavily selected for and globally broadcast and especially importantly rehearsed (when you report to yourself, that information becomes very long-lasting and stable), even if the signal wasn't globally broadcast before it became reported. Essentially the language faculty allows some signals a free-win, and so whatever draft of a thing is accessible to the language faculty will be the draft that people feel was the true experience, as when prompted on what they experienced as it will be the one they report, and thus the one they remember and continue to think about.
When someone's streams of input is 'probed' (they ask themselves or are asked what there experiences were), a single most-popular draft becomes 'precipitated', selected for above all others, canonized, in a sense. This provides a 'time line; a subjective sequence of events from the view of an observer'. This time line allows us as scientists to compares peoples subjective timeline to other timelines such as the timeline of neurological processing which may or may be match up with the subjective one.
A very core idea of Dennett's is that there is no facts about experiences independent of probes, since without probing there are no privileged drafts in the brain, only many different ones. Probing is important because as mentioned it involves a kind of sustained rehearsal and global distribution of contents that language facilitates.
However autostimulation exists even without a functioning language faculty, and is constantly everpresent in the brain, meaning some contents are still being 'rehearsed', providing a sort of sublinguistic time line. Essentially in the brains of humans who lack the language faculty but retain autostimulation, they are still sub-linguistically probing themselves (or rather, continuing to act and think under the influence of probes on neural habit: with some contents rehearsed and globally distributed as if they were in language).
Dennett's original view was that conscious states were nothing more than popular states, since then he's seemed to change his mind a little referencing meta-representations as essential for consciousness[2]. Regardless, he basically claims that whatever contents are being globally rehearsed and processed are the contents of your consciousness. He calls the contents of consciousness the contents of the Joycean machine (which he calls a serial virtual machine loaded onto parallel hardware), as it contains only those contents which are being rehearsed in a special, global way facilitated by language.
Most of the content of the Joycean machine is non-linguistic, but it is the influence of language that facilitates the existence of the machine.
Animal (un)consciousness
Dennett gives two core reasons why he sees human consciousness as so profoundly different from animal consciousness (+ mostly from young human baby consciousness too), and thus why he doesn't see animal suffering as very serious.
Selfless
The first is that non-human animal minds lack a self model. They have no concept of 'I' woven in their brain, and because of this they have no narrative centre of gravity, no meaningful model of themselves as an individual in the world. They are perpetually disassociated. Dennett makes reference to disassociated traumatised children who report no longer or mostly no longer feeling pain due to their disassociation, using this as an indication that animals might be the same way as they never even develop a self to dissociate from in the first place. [3]
He then says in another article, for 'body protecting states' to matter, there must be an 'enduring, complex subject for whom they matter because they are a source of suffering'. In thinking about it this way, bodily damage induces a whole extra set of judgements / interpretations unique to humans, judgements that in Dennett's view constitute suffering. Put another way, the experience of pain to a creature with no narrative self is so fundamentally different in his view it isn't really comparable to humans, where mostly everything we recognise as suffering exists only because of the narrative self.[4]
Unified mind
Daniel Dennett consistently claims that non-human animal minds lack unity, as although they processes information, without the 'timelines' provided by autostimulation, there is never any point where the organism as a whole comes to perceive anything, there are only parts of organism's brains working with different information.
He mentioned his precise beliefs and predictions here, saying that long distance reverberant / re-entrant interactions have to be laid down by language use, and that without these connections there is no functional unity to the nervous system.[5]As a result, strictly speaking, there can be nothing 'it is like' to be a cow, for example, since there is no cow-system, only different sub-systems within the cow with their own shifting loci of activity.
He also seems to argue against the functional unity of representations in non-humans minds as well. In one example he points out that if you train a rabbit to be fearful of something that its seeing with its left eye, this won't generalise so when the rabbit sees the cue on its right it won't be scared. 'What's it like to be a rabbit? Which one, the left or the right?' he says, implying their disjointed reactions underlie a disjointed awareness of the world. [6]
In another example he points out research showing that snakes use 3 separate modalities to catch mice, always using the visual system to strike and locates it via smell, and then devours it via tactile information. Notably, according to the article even if the mouse dies right in front of the snake it will still follow the smell trail anyway. They also point out that this behaviour is especially some snakes like boas and pythons, where even though the snake 'must have ample proprioceptor-y information' about the location of its prey, it searches for it in its grasp with only its scent anyway. From this he claims the snake lacks a central representation of the mouse its chasing.
Presumably then language learning in his view doesn't just allow for a unified conscious mind, but also allows reliable central representations of objects in one's world, mind and body (it facilitates general schema building).
Pretty much all this information comes pretty directly from his book Consciousness Explained, but you can also find it mentioned it most of his writings.
It seems like an additional power of language for Dennett is the power of meta-representation, as language allows you to work with your own representations.
I'll note that dissociated children seem to show a lack of response to pain primarily because of a lack of attributing pain to a body schema.
If you're interested in this it might be worth reading this report, where Kaj seems to report a pretty significant dissolution of the narrative self while retaining otherwise normal functioning, finding a significant drop in negative affect and most of all aversiveness of negative experiences. There was a dissolution of craving and a breakdown of a 'subjective glue' binding things together. I don't really know what this means but do wonder if this could imply, contrary to the Buddhists, that animals are actually more protected from cravings than humans are and thus do suffer less as they become less entangled in things.
In regards to global workspace theory, my best guess is that he believes the all-or-nothing behaviour associated with the global workspace (where a consensus is reached within the workspace on interpretations of input via reverberatory / re-entrant activity) is caused by the influence of language, whereas non-human animals only have contents that are in varying stages of globally distributed without any of these being privileged. This tracks with his strong association between autostimulation and re-entrant connections and the very high importance of long distance re-entrant connections to the functioning of the global workspace.
I found this example very unconvincing, as its not like the rabbit understands why its been conditioned into being fearful in the first place, so it easily could have a unified mind but simply feel a involuntary jump of fright at seeing it one eye but not the other.