Talk:Perceptual control theory

Psychotherapy and MOL

Should the information now under "Psychotherapy" be put instead in the separate Wikipedia article on the Method of Levels (with a link from here to there)? --Gary — Preceding unsigned comment added by Gary Cziko (talk • contribs) 21:06, 25 July 2005 (UTC)

I don't think so. PCT Psychotherapy is a more general concept than the Method of Levels. It is possible that a person in therapy will change in more ways than resolving conflicts. For example, a person may learn to perceive differently, want differently, feel differently, and act differently. Changes can take place anywhere in the control loop. --David — Preceding unsigned comment added by 209.204.124.231 (talk) 02:44, 27 July 2005 (UTC)

This article look like a commercial. Major editing required. —Preceding unsigned comment added by 83.130.10.63 (talk) 00:24, 25 March 2008 (UTC)

Article is a complete joke. Reads like a horribly written and inaccurate 6th grade textbook. Either needs to be completely edited or deleted. The theory was developed by William T. Powers, "a maverick scientist." AKA hack. He has no institutional affiliation which should be a big red light considering he seems to be the only person backing this theory. —Preceding unsigned comment added by 137.165.249.79 (talk) 05:31, 14 January 2009 (UTC)

Actually you are deeply mistaken in this evaluation. Several cyberneticists, behavioral scientists, and theoretical biologists have arrived at similar conclusions albeit stated in different terms -- Ernest Von Glasersfeld, Ludwig Von Bertlanffy, Norbert Weiner, W Ross Ashby, Humberto Maturana, Fransisco Varela, Heinz Von Foerster, Gregory Bateson all of whom were associated with respected institutions and who have written numerous publications in the behavioral and biological among other sciences in numerous scientific journals, not mentioning that they held respectable positions in respectable institutions. A lot of this work has thoroughly been picked up on in sociology, anthropology, and cognitive science, for example in the works of Tim Ingold, Niklas Luhmann, Andy Clarke, Soren Brier, Evan Thompson, among others. It appears as if PCT is an independent formulation of many of the similar ideas of these other researchers, however not likely since they all reference cybernetic notions in some form or another in all of their works. And if you look at the literature Powers actually wrote, you will see him mention some of these names as he compares and contrasts his notions with theirs. His works were also read and appreciated and favorably reviewed by other behavioral scientists, of course not most because he was deeply challenging the behaviorist orthodoxy. It just seems that cybernetics outside of the engineering domain of automatic control systems theory has been ignored by most of mainstream academia in the united states as historically cybernetics was abandoned for more "concrete" research in artificial intelligence and computing. Detailed social histories have been written about this state of affairs by sociologists, which I could find if you are interested in seeing that these ideas are not just the result of a lone "hack" working in isolation. DivisionByZer0 (talk) 09:05, 14 March 2009 (UTC)

DivisionByZer0, this is cognitive psychology, not sociology. Your extreme name dropping isn't the slightest bit convincing. Saying all sorts of bodies of work have related "formulations" of "similar ideas" isn't real evidence. If you have real evidence that modern psychologists are using Powers' theories--or some salient derivation thereof--in cognitive science, then show that evidence. (Ironically, I actually have some evidence, although weak.)

As for the content of the article itself, while Powers and various theoretical descendants may come off as hackish, I recall seeing Powers cited by Michael_Tomasello in a 2005 paper [[1]]. I'm not saying this mention legitimizes any of the specific content on this page, but simply that PCT being covered well on Wikipedia, with NPOV, without all this COI, would be generally useful. 76.115.3.200 (talk) 09:27, 19 August 2009 (UTC)

Begging to differ, this is presented as a science of the behavior of living things, and in point of fact there is a solid literature on PCT and sociology, e.g. Mc Phail, Clark, The Myth of the Madding Crowd, New York, Aldine de Gruyter, 1991. Cognitive psychology holds that stimuli, mediated by 'cognitive maps', cause behavior—a very different beast. Bn (talk) 19:02, 19 December 2010 (UTC)

Suggestions for improvement

The Cleanup tag gives as reason "The article is difficult to understand". It would help if the person who applied the tag could identify specific sections that are written in an unclear way. Bear in mind that many articles about fields of science are hard for newcomers. Of course good science writing for the lay public aims to overcome this, but the reader is expected and required to do some work. Readers who are expert in some fields may not realize the extent to which they are newcomers to this.

There could be a section on each area of application. These may become substantial enough to merit being split off into a separate article, as with MOL. For example, there should be a section on PCT and sociology, which should be put in relationship with Control_theory_(sociology), and then the disconnected subsection of references to PCT and sociology could be properly handled.

The article Control theory (sociology) has some things that seem a bit garbled to me, quite apart from not mentioning PCT. For example, the paragraph that begins "Control Theory, as developed by Walter Reckless in 1973, states that behavior is caused not by outside stimuli, but by what a person wants most at any given time" concludes with the frankly incoherent sentence "So, basically, if you have strong social bonds to positive influences, deviant behavior is less likely than someone who has no family or friends." Look at the talk page at Talk:Control theory (sociology) to see how a series of three editors started to deal with this a couple of years ago—the last of them deleted a plug for 'choice theory' in October 2009, and they agreed that this Walter Reckless's ideas amount to a Hobbsean "human nature is inherently depraved" approach to criminology which is unrelated to negative feedback control theory.

More needs to be said about the distinction from engineering control theory in the section that has been added for that. A number of articles are related but assert some of the familiar misconceptions about negative feedback control, e.g. lag time. These include Control theory, Feedback controller, and Feedback. Bn (talk) 01:36, 9 January 2011 (UTC) The introduction mentions this distinction, and nothing should be in the introduction unless it is taken up in the main body of the article. Bn (talk) 17:34, 9 September 2012 (UTC) Bn (talk) 04:29, 22 September 2012 (UTC)
Bn (talk) 02:01, 1 October 2012 (UTC)

Removing cleanup tag

On 4 February 2006, A cleanup tag was inserted (sans required parameters) by Scm83x.
Smackbot added a date in June 2008.
On 20 May 2012, IP 68.55.112.31 added the reason "The article is difficult to understand." Because the syntax was not correct it did not display. On the same date, the same IP added and then reverted a further reason "Also, the article lacks sufficient information for the reader to evaluate the level of scientific support for the theory." The edit note says:

(Propose content for the mandatory "reason" parameter to the "cleanup" tag. Please feel free to change the "reason".)

Recently, I fixed the syntax so we can see the reason displayed.
Today, I put a note on the talk page for Scm83x stating intention to remove the tag. (The talk page for the IP says it's a work computer that is blocked for editing until April 17, 2012.)
The page has been greatly improved. If there is no objection, I will remove the cleanup tag next week.
Bn (talk) 05:49, 26 September 2012 (UTC)
[Done.] Bn (talk) 02:07, 1 October 2012 (UTC)

State-of-the-art

This text need objectivity and context.

1. A comparison with other state-of-the-art models on perception
2. If this is a "model" as stated in the introduction, make it specific. What are the model's constituents? What are the model parameters? What is it a model of? Which different variants of the model exist? How can the validity of the model be tested?
3. Use existing terminology, there has been written a lot about sensorimotor prediction to mention one example. This shows also that you know the alternatives (and why they are not good enough)

To the author: Do not see this too much as a criticism. Other cognitive models seems to have the same lack of rigour, as there are a.) global workspace theory (Baars), b.) society of mind (Minsky). See for examples of a little bit more specific models c.) copycat (Hofstadter), and d.) HTM. Also compare with models on the cognitive architectures wiki page. I would recommend to make use of well established notions such as Bayesian networks, Shannon information, entropy, sensorimotor mutual information, free-energy (Friston), empowerment (Klyubin), information-to-go (Polani), etc. Andy (talk) 20:00, 23 April 2012 (UTC)

This article has still not improved. The 'mathematical' section is not worthy of its name. Moreover, it is well known that minimizing error leads to a system that does not exhibit exploration, but only exploitation. Moreover, it eludes me why in this case the reference signal "r" coming from "another part in the organism" is so different from an external signal. If so, it needs to be specified how this signal is generated. For now it really comes across as a crackpot article, worthy only of deletion. Anne van Rossum (talk) 18:11, 23 December 2013 (UTC)

Good catch re the vague language about the source of the reference signal. However, you should try to get at least a little exposure to the relevant literature before shouting "crackpot". Bn (talk) 23:08, 28 July 2014 (UTC)

Conflict of Interest

Gary Cziko has personally published papers about PCT, and created much of the content on this page, and edited this page to include links to his papers, to boot. He also created the Method of Levels page. While the lack of NPOV is obvious, this COI might not have been. I don't know a lot about Wikipedia policy, but I'm going to make a good faith assumption that Gary Cziko was just trying to share his knowledge, despite it reading like an advertisement. I will post on his talk page that he really shouldn't be doing this. After a few minutes of reading the COI page, I figure this is the best unofficial course of action on my part. Hopefully a real Wikipedian can include any formalities, like the COI template thingy, if warranted. 76.115.3.200 (talk) 09:27, 19 August 2009 (UTC)

This entire article looks like an advertisement. I just don't have the guts to delete half of it. Some of it obviously has to be removed while other parts have to rephrased. Yelin (talk) 04:55, 17 November 2009 (UTC)

I have rewritten the preamble, and included references to the reports of NATO groups of which I have been a member, as well as to a Special Issue on Perceptual Control Theory of the International Journal of Human-Computer Studies of which I was the Editor. Since I did not include evaluative comments on these, I'm not sure whether my mention of them constitutes conflict of interest, so I declare it here. Mmt (talk) —Preceding undated comment added 04:35, 15 May 2010 (UTC).

I would have added the references to Cziko's MIT Press books if he hadn't. They are valid, reliable sources. No conflict of interest. That was the original complaint. Nor do I see any conflict in the other references. Nor is there an issue of notability, there are many mainstream references and publications, so it is notable enough for its own article. Mention of PCT within an article about, say, the Skinner box would require reliable sources that recognized the relevance of the one to the other, but notability is not an issue for that sort of cross-inclusion even if this standalone article were challenged on notability grounds (which it is not).

Unsigned IP comments tend to have reduced credibility. Please sign your comments. Bn (talk) 16:54, 1 October 2010 (UTC)

Non-mathematical?

The opening statement of the article states that PCT is non-mathematical. That's at best misleading, but more likely it's just wrong. While you can understand much of the theory without working through the math, it does have a mathematical basis in control theory (see the appendix in Powers' 1973 book) and the computer simulations are built on that math. Bill Benzon (talk) 18:01, 24 May 2012 (UTC)
This has been fixed by reverting an edit of the intro and by adding a mathematical section. Bn (talk) 02:08, 1 October 2012 (UTC)

Maurice Merleau-Ponty

An unsigned IP made this the opening sentence of the History section:

Maurice Merleau-Ponty, influenced by the work of Gestalt psychologists, laid the foundations for PCT in his 1942 The Structure of Behavior,<ref>The Structure of Behavior, A.L. Fisher (trans.), Boston: Beacon Press, 1963.</ref> although he remains mostly unacknowledged.

This would require citations showing two things:

1. That Merleau-Ponty understood negative feedback control as the explanatory principle for behavior.
2. That his work in fact influenced the development of PCT.

Seeing no evidence for these prerequisites either in the WP article on him or in the literature of PCT, I have removed this passage. If such citations can be provided, we should find an appropriate place to mention him.
Bn (talk) 01:09, 28 October 2013 (UTC)

Reverting uninformed tag by anonymous IP

An anonymous IP user inserted an ad tag, with the following edit comment: "This article currently reads as an extended advertisement for a marginal concept that is not accepted by most psychologists."

This is a claim that PCT is a "fringe" topic relative to "mainstream" psychology. A challenge by WP:Fringe applies only in context of WP:Undue when a "fringe" topic is mentioned within an article about a "mainstream" topic. It does not apply to a standalone article on a topic whose notability is unquestioned. If that user wants to challenge the notability of PCT, he or she should please step forward with appropriate discussion here. A cite tag is appropriate to mark passages that seem controversial if they lack cited support from reliable sources.

I am removing the ad tag, awaiting that further discussion.

Bn (talk) 15:43, 12 December 2019 (UTC)

Slate Star Codex RS

user:Greyfell removed a ref to a review by Scott Alexander [last name omitted] on ground that his blog was not a reliable source. It is considered notable by WP standards, as witness the Slate Star Codex article. The content of that article, and of the RS publications on which it is based, indicate that it is widely regarded as a reliable source. Bn (talk) 23:35, 24 November 2020 (UTC)bn

No, it is not a reliable source, since per our own article, it is a blog. Blogs fail WP:SPS. It may be "widely regarded" by some readers as reliable, but this is not the same as having a reputation for accuracy and fact checking, which is the base-line standard for WP:RS. It might be usable with clear attribution, but that would generally also require WP:IS coverage indicating why Alexander's personal opinion is encyclopedically significant. Grayfell (talk) 23:40, 24 November 2020 (UTC)

Original research and neutrality issues

@Bn: Hello. I originally found this article by reviewing external links to the Slate Star Codex blog, which had a minor spam issues previously. After your revert attracted my attention to it again, I have been looking over it more closely, and I see a number of serious problems.

I preface this by saying that identifying some issues is not an endorsement of the rest of the article, it's merely intended to be a starting point.

The article uses a very positive tone to describe PCT, which fails WP:NPOV. Having the lead imply that this model is efficacious, using terms like "extremely high", "vanishingly rare", using the word solution... these fail WP:TONE and potentially WP:PEACOCK. None of these terms are forbidden, but they need to be evaluated in context, and in context, they are non-neutral.

The first subsection, #The place of purpose (intention) and causation in psychology appears to be original research. Citing a source from 1924 is just one red flag that this section was written by a knowledgeable editor attempting to provide an overview they personally thought would be helpful. Obviously this is a reasonable starting point, but it's not really enough. The issue is that we need to summarize reliable sources in almost all cases. This includes context. For almost all content, sources must directly mention "perceptual control theory" or an uncontested synonym. Any sources which do not use this term should be supported by sources which directly explain the connection to this topic. A very brief summary of background issues based on unrelated sources may be acceptable, but it is never ideal. When background information is necessary, one good solution is to start with links to other Wikipedia articles. Category:Mental processes and Category:Concepts in the philosophy of mind might be helpful.

Also, the title of this subsection is unwieldy, to say the least. Sections should be clear and simple. It will only confuse readers to try and explaining secondary topics like this. This article is not about "the place of purpose (intention) and causation in psychology" nor does PCT have a monopoly on explaining this "purpose". It is therefor not appropriate for the section to bypass other explanations of this based on WP:OR. Start with what reliable, independent, secondary sources are saying about PCT and work backwards.

I could go on, but perhaps this should go to a noticeboard to that other knowledgeable editors can weigh in. Grayfell (talk) 00:36, 25 November 2020 (UTC)

I've made a start by subordinating the "place of purpose" discussion within the history section, because it really does have to do with historical roots and does not belong as a kind of preface. 'terms like or replacing "extremely high", "vanishingly rare",' etc. or equivalent can be quoted from RS or cited to passages in them to make clear that they are not the product of some editor's enthusiasm. I've asked others to weigh in. Further developments may be delayed depending on ongoing circumstances.

Bn (talk) 19:54, 2 January 2021 (UTC)

Methodology

The section states that "a generative model is built which replicates the data observed for individuals with very high fidelity (0.95 or better)". What is meant with "0.95"? A correlation coefficient? A coefficient of determination? Jwdietrich2 (talk) 10:39, 24 December 2022 (UTC)

More on neutrality issues

I don't usually pay attention to this article, but I just happened to look at the most recent edit by Jwdietrich2 and one sentence in it raised a red flag for me; it says of PCT:

Unlike simplifying theories including behaviourism and cognitive psychology it sets out from the concept of circular causality.

A similar claim appears in the lead section:

It differs fundamentally from theories of behaviorism and cognitive psychology which model stimuli as causes of behavior (linear causation).

And elsewhere in the article:

There follows from this stance the assumption that environmental events (stimuli) cause behavioral actions (responses). This assumption persists in cognitive psychology, which interposes cognitive maps and other postulated information processing between stimulus and response, but otherwise retains the assumption of linear causation from environment to behavior.

All of this seems to be accurately parroting the narrative in the cited article by PCT proponent Richard Marken, but I doubt the narrative is true, because there is cognitive psychology that does not fit this narrative. Even Cognitive psychology § Criticisms notes the diversity of models as a weakness of cognitive psychology, quoting psychologist John Malone as saying: "there are many, many varieties of cognitive psychology and very little agreement about exactly what may be its domain".

I just did a quick Google search for circular systems cognitive psychology, and the top result is a 2015 article in Frontiers in Psychology by David Vernon et al. that includes a survey of some cognitive psychologists who use circular causal models such as the work of Lawrence W. Barsalou et al. on situated cognition and embodied cognition, Mark H. Bickhard et al. on interactivism, and others. To see whether this work is reflected in cognitive psychology textbooks, I checked a random textbook, Cognitive Psychology and Its Implications (9th ed.) by John R. Anderson, and it does have a section on the embodied cognition of Barsalou et al. with a quotation from Esther Thelen that begins: "To say that cognition is embodied means that it arises from bodily interactions with the world and is continually meshed with them." The section's summary says: "The embodied cognition perspective emphasizes that meaning is represented in the perceptual and motor systems that we use to interact with the world." Furthermore, Wikipedia's article on embodied cognition has a diagram that clearly shows it as part of cognitive psychology (and other fields): File:The scope of embodied cognition 06.10.2021.png.

It appears to me that the phrases quoted above need to be modified to be more qualified and to indicate exactly which model(s) in cognitive psychology the PCT proponents are against. Otherwise the quoted phrases make false overgeneralizations about current cognitive psychology. In particular, the phrase Unlike simplifying theories including behaviourism and cognitive psychology seems extremely loaded and non-neutral. (As if PCT wasn't a simplifying theory! Of course it is!) Biogeographist (talk) 20:52, 5 April 2023 (UTC)

I think you are right. Thanks for pointing out this problem and for rephrasing the problematic parts of the article. Jwdietrich2 (talk) 17:47, 12 April 2023 (UTC)

Notifying Grayfell, who was the last editor to notice neutrality issues in this article, in case they wish to add feedback. Biogeographist (talk) 20:52, 5 April 2023 (UTC)

References