[Philosophy of Social Cognition] Twelfth Meeting

Martyna Meyer martyna.meyer at univie.ac.at
Wed May 31 18:19:40 CEST 2023


Dear all,

I hope you're doing well!

Many thanks to the participants of yesterday's meeting, I think we 
managed to dissect the paper very thoroughly.
A special thank you to Moritz for the impromptu presentation on Bayesian 
Predictive Coding.

In this email, I'm sending you:
I. The paper and the meeting details for next week
II. A *whole bunch* of notes/comments/resources regarding our last 
meeting (thank you so much: Bailey, Moritz, Jonas, Flavia, Felix)

- - - - - - - - - - - - - - - - - - - - - - - -

I.  For the next session, we are reading:

(Attention: This text is a new addition to the syllabus. I changed the 
website information <https://socialcognition.phl.univie.ac.at/syllabus/> 
today.)

Palmer, C. J., Seth, A. K., & Hohwy, J. (2015). *The felt presence of 
other minds: Predictive processing, counterfactual predictions, and 
mentalising in autism*. /Consciousness and Cognition/, /36/, 376-389. 
https://doi.org/10.1016/j.concog.2015.04.007
You can download it here 
<https://www.sciencedirect.com/science/article/pii/S1053810015000847?via%3Dihub> 
(free access).

***
If you're super interested in the topic and want to be extremely 
well-prepared, you can check this out, too:
- Smortchkova, J. (2022). Face perception and mind misreading. /Topoi/, 
41(4), 685-694. (access here 
<https://link.springer.com/article/10.1007/s11245-022-09823-z>)
- Smortchkova, J. (2020). Does empirical evidence support perceptual 
mindreading? /Thought: A Journal of Philosophy/, 9(4), 298-306. (access 
here 
<https://www.researchgate.net/publication/345941160_Does_empirical_evidence_support_perceptual_mindreading>)
- Varga, S. (2018). Toward a Perceptual Account of Mindreading. 
Philosophy and Phenomenological Research, 100(2), 380–401. (UniWien 
access here 
<https://libkey.io/libraries/2327/articles/233785982/full-text-file?utm_source=api_1269>)
***

As usual you’re welcome to join our session online (Zoom link 
<https://univienna.zoom.us/j/65514918078?pwd=cVZTd2Ivb09uSUFVNTZORWFIOTA4UT09>) 
or in person, at *NIG* (room *3B*, third floor).
We’re starting at *6:30 CET*. The next meeting is on *Tuesday*, *June 
5*, 2023.
Newcomers are absolutely welcome*. Please join us!

*Previous knowledge is not necessary to participate, but if you have a 
wish to briefly orient yourself in the debate, this is a good intro text:
Fenici, M. (2017). Rebuilding the Landscape of Psychological 
Understanding After the Mindreading War. Phenomenology and Mind, 12, 
Article 12. https://doi.org/10.13128/Phe_Mi-21113

- - - - - - - - - - - - - - - - - - - - - - - -

II. (a) Resources from Bailey on games:

Hey! Here’s some of the stuff on games.

The first book I had in mind was

Suits, Bernard. The Grasshopper: Games, Life and Utopia. New York: 
Broadview Press, 2005.

Unfortunately, it doesn’t seem to be in any of the university libraries 
(!), although a lot of responses to it are. Suits provides a definition 
of games as the taking up of artificial ends, which are only pursued for 
the sake of play. His big (philosophical) move is to argue that that’s 
basically what life is. As you can see from the title, it’s somewhat 
oriented towards a  popular audience. That shouldn’t deter you, though, 
if you’re interested: the underlying arguments are (to me anyway) still 
quite good.

This analysis has been extended to *actual* games in recent years by C 
Thi Nguyen. Most of Nguyen’s work goes more towards aesthetics than 
practical reason: he’s interested in what the constitutive features of 
games as an art form are. His answer is basically Suits’: they 
crystallize agency.

For Nguyen,  see

Nguyen, C. Thi. “Games and the Art of Agency.” The Philosophical Review 
128, no. 4 (2019): 423–62. https://doi.org/10.1215/00318108-7697863

Nguyen, C. Thi. Games : Agency as Art. New York, NY: Oxford University 
Press, 2020.

There’s a passage on “stupid games” in the book —- games played where 
the whole point is to fail — that I really like. I think it’s in the 
first chapter.

Best
Bailey

- - - - - - - - - - - - - - - - - - - - - - - -

II. (b) Resources from Moritz on schizophrenia (as promised to Flavia) & 
(/vanilla/) intro to Predictive Processing:

Here is the schizophrenia paper, which is actually more generally about 
psychosis but discusses schizophrenia:
and I was wrong, there is actually still research going on and people 
are publishing about it even in clinical settings:

https://www.sciencedirect.com/science/article/abs/pii/S1053811918305007
https://www.cambridge.org/core/journals/bjpsych-advances/article/innovations-in-the-psychopathology-of-schizophrenia-a-primer-for-busy-clinicians/7FB62E9ACFE88AE026AB3899587E39E0 


And second is PP for philosophers which is a great summary and probably 
fits the reading club!

Best,
Moritz

[Moritz sent two attachments, you can find them below:
"Vanilla PP for Philosophers" & "The Predictive Coding Account of 
Psychosis" papers]

- - - - - - - - - - - - - - - - - - - - - - - -

II. (c) A lovely follow-up question from Jonas about Predictive Coding:

[...] for the second half of the session [I] just pondered the following 
question which then never felt NOT either out of place or too uninformed 
to ask:
For me, predictive coding can't really actually explain new things that 
"standard" approaches, which assume input x with following computation 
of and reaction to x, can't. I understand it only on a computational 
level of neuronal mechanisms, eg,

classical: There's a set of {input that happens and input that doesn't 
happen}. The "cognition" then goes on to process (input that happens) 
and outputs accordingly.

predictive coding: There's a set of {input that is predicted to happen 
and input that is not}. The "cognition" then goes on to process the 
(input that was not predicted to happen).

Therefore, both concepts work on very similar informational bases, 
although kind of inverted with respect to each other.

"Cognition" here not referring to some philosophical, eg, sandwich, 
model, but to computations following informational output from direct 
sensory "perception" (as in: hair cells get depolarized by mechanical 
forces etc - no interpretation here yet). Therefore, predictive coding 
does nothing more (and nothing less) then to invert the mechanisms of 
brain cells firing in a more efficient way, if you will, because it 
allows for quicker processing (most relevant computations happen BEFORE 
an event) as well as for more efficiency (don't process the 100 things 
that are always similar, but the 3 that are new in every situation). But 
it does not have explanatory reach to other issues, like explaining mind 
reading in terms that "classical" approaches couldn't. For me this is 
two different layers of analysis.

Why is that? It's obvious, even in standard approaches, that after the 
sensory Input, a whole lot of computations kick in. What may be termed 
"mind reading", either by explicitly theorizing or by implicitly 
simulating as basis would also be based on these computations, and those 
theories make (to my knowledge) very little assumptions as to the 
biomechanical reasons that could lead to the phenomenon of 
"understanding other minds". The perception of explicitly theorizing may 
arise both in "reaction" approaches as well as in "prediction" ones, no? 
We don't know how that perceptional quality arises anyway. (and of 
course simulationist approaches do offer the sharing of neuronal 
representations as basis for understanding - but this could just as much 
signal *prediction* of how one should feel / how the other feels)

This goes also for Andreas' example with the "looking ahead of the 
moving phone" - to me that was not really relating to predictive coding 
vs classical approaches, but rather could be explained in both terms. In 
predictive coding it's clear: You don't actually follow the phone, but 
rather it's predicted trajectory, and react, should it diverge. But also 
in classical approaches, you would get the input of the phone moving in 
a circular motion, leading you to infer that it stays on this 
trajectory; and this would - even in this reactionist approach - 
constitute an explicit prediction that would then have to be tested. In 
my mind, automatic computations based on previously experienced events 
that formed patterns based on regular occurence obviously impact 
computations in both cases (reactionist and predictive) - things as 
"simple" as walking prove that, without having to compute muscle 
movements for every step explicitly, it follows an automatized and very 
self-regulating pattern of muscles movements. Same case for catching the 
ball that was mentioned. Therefore, I would argue that classical 
approaches would never say: There is no prediction (on an automatic, 
maybe even probabilistic, basis, in some cases) - obviously there is! 
They may just (implicitly) differentiate between prediction as (1) 
abstract concept and as (2) possible neuronal computation mechanism. 
Therefore, for me, the difference to predictive coding could only become 
visible on the neuronal layer: In the case of the phone, is it the 
staying-on-trajectory that triggers most neuronal spikes, or is it the 
diverging?

And of course, predictive coding also introduces this concept of model 
generation and then probabilistically testing and updating these models 
against and with "real input"; But is that concept so irreconcilable 
with standard approaches? Here you may get input x, then "theorize" or 
"cognitize" different possible options as to what it may mean and what 
may be best, and then execute the "logically superior" alternative. 
Maybe the difference here lies in more deterministic classical 
assumptions vs probabilistic ones in predictive coding?

[...] in short,*for me predictive coding is not really a novel theory to 
explain cognitive phenomena, but rather a novel theory to allow for more 
efficient computation on the layer of neuronal firing on all hierarchies 
of the brain. *

Nonetheless, yesterday we talked a lot about also what it could explain 
and what it couldn't. I really would be curious - am I just not too well 
read in the literature (I am not), or may there be a point to this 
perceived fallacious intermingling of concepts (computational layer - 
reaction vs prediction, VERSUS conceptual layer - understanding others, 
read minds, simulation, ..., whereas both layers do not have explanatory 
influences on each other, because it is neurochemical VERSUS abstract 
concepts, whereas the latter may or may not directly relate to actual 
mechanical brain function?)

I guess it might have been nice to ask this yesterday, but it just 
started to grow in my head mid-way and it wasn't fleshed out enough to 
ask sensibly.
So, erm, here is this wall of text now, because I'm CURIOUS

Best,
Jonas

- - - - - - - - - - - - - - - - - - - - - - - -

II. (d) A paper from Flavia:

Hello all!

This interested me:

Di Paolo, E., Thompson, E., & Beer, R. (2022). Laying down a forking 
path: Tensions between enaction and the free energy principle. 
Philosophy and the Mind Sciences, 3. 
https://doi.org/10.33735/phimisci.2022.9187

It’s quite a recent paper. Maybe the rest of you would like to check it 
out too. Its not SocCog oriented, but related to some overlaying 
questions of yesterday.

Lovely weekend everyone!
F.

- - - - - - - - - - - - - - - - - - - - - - - -

II: (e) A meme from Felix [in the attachment]
(Yes, we are making social cognition memes now! It makes me so happy)

Many, many thanks for all the resources and comments.
I'm looking forward to seeing you next week. Everybody is very welcome 
to join.

Best,
Martyna
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.univie.ac.at/pipermail/socialcognition/attachments/20230531/6b1e6d19/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Vanilla PP
Type: application/pdf
Size: 674581 bytes
Desc: not available
URL: <https://lists.univie.ac.at/pipermail/socialcognition/attachments/20230531/6b1e6d19/attachment-0002.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: PP of psychosis
Type: application/pdf
Size: 518590 bytes
Desc: not available
URL: <https://lists.univie.ac.at/pipermail/socialcognition/attachments/20230531/6b1e6d19/attachment-0003.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: TheSandwichModel.jpeg
Type: image/jpeg
Size: 73401 bytes
Desc: not available
URL: <https://lists.univie.ac.at/pipermail/socialcognition/attachments/20230531/6b1e6d19/attachment-0001.jpeg>


More information about the SocialCognition mailing list