Talk:ELIZA effect

This article is rated C-class on Wikipedia's content assessment scale.
It is of interest to the following WikiProjects:

Computing

	This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.ComputingWikipedia:WikiProject ComputingTemplate:WikiProject ComputingComputing articles
???	This article has not yet received a rating on the project's importance scale.
	This article has been automatically rated by a bot or other tool because one or more other projects use this class. Please ensure the assessment is correct before removing the `\|auto=` parameter.

Philosophy: Logic Mid‑importance

This article is within the scope of WikiProject Philosophy, a collaborative effort to improve the coverage of content related to philosophy on Wikipedia. If you would like to support the project, please visit the project page, where you can get more details on how you can help, and where you can join the general discussion about philosophy content on Wikipedia.PhilosophyWikipedia:WikiProject PhilosophyTemplate:WikiProject PhilosophyPhilosophy articles

Mid

This article has been rated as Mid-importance on the project's importance scale.

Associated task forces:
/
	Logic

This article is based on material taken from the Free On-line Dictionary of Computing prior to 1 November 2008 and incorporated under the "relicensing" terms of the GFDL, version 1.3 or later.

Wiki Education Foundation-supported course assignment

This article was the subject of a Wiki Education Foundation-supported course assignment, between 20 January 2021 and 21 April 2021. Further details are available on the course page. Student editor(s): Kellychen927. Peer reviewers: Alexakapoor, Oliviaolcott.

Above undated message substituted from Template:Dashboard.wikiedu.org assignment by PrimeBOT (talk) 19:57, 16 January 2022 (UTC)[reply]

Cross purposes

"... there is nothing magic about the symbol "+" that makes it well-suited to indicate addition..."

Actually there is something, even if not magic; The cross is one of the most ancient symbols used by mankind, probably in use long before writing. It is associated with the idea of two things meeting, most notably earth (horizontal plane, -) and sky (veritcal, |) or the meeting of weapons in combat (x), so the meeting of numerical values is pretty close. A modified version of the symbol, the T cross (only associated with the letter t in the cultures that use it, before modern communications and the spread of english that is) implies one componenet supporting another, thus when the horizontal component is placed on the vertical the result is a structure that has a different identity than either of its parts. An example would be wooden beams in buildings, bridges or other constructs.

I should mention that i have no knowledge of the actual history of the cross as a mathematical symbol, so the above text is more an opinion than a report of facts. This opinion however does come from the study of other subjects (not related to mathematics) like history and anthropology.

It was made up in the Middle Ages, after words had been used. lysdexia 01:50, 13 Nov 2004 (UTC)

Worse, the + symbol is just a whittled-down P.

"A little learning is a dangerous thing."

— Alexander Pope, An Essay on Criticism

We are watching myth-making in action. The original poster has no knowledge of the subject (as he admits), but instead of feeling constrained by real-world facts, generates a story made up out of his own head. If lysdexia hadn't responded, how many readers would have taken this fairy tale as the truth? It's not as if the history of mathematical symbols is shrouded in darkest mystery; the poster just didn't bother to check before making his statements. B00P 16:12, 9 May 2006 (UTC)[reply]

I don't quite get how this is related to ELIZA, the chatbot. Is it because ELIZA just throws terms back at you without itself understanding what they mean? --AaronW 18:24, 17 Mar 2004 (UTC)

In my opinion, it is due to the fact that humans talking to ELIZA interpret what it says as intelligent. --Parmentier 08:33, 24 Jan 2005 (UTC)

Deletion?

I'd vote for rewriting or deleting this article. It is not clear from it that this is computer-slang. Rather, it seems to imply that this vaugely-defined 'effect' originated with ELIZA, which is completely untrue. The whole field of Semiotics is devoted to studying these relations, and it is quite old. This is just a (very rare) computer slang term which seems to encompass the whole field. This would be the wrong place to try to give a full description of semiotics and semiotic relationships.

Yes, this article needs a rewrite. I don't wote for deletion, since the word is a part of the computers history.

It is hard to find any online information about the Eliza Effect, since the same definition found here has been copypasted all over the web. However, the book "Hamlet on the Holodeck" defines the Eliza Effect as "attributing greater intelligence or intentionally to a machine than it possesses".

I think this is much more exact than all the talk about symbols. --Kasper Hviid 10:22, 6 August 2005 (UTC)[reply]

Revamped ELIZA effect page

Alba recently rewrote the page from scratch and the result is clearly an improvement. However, I'm concerned about a few things:

There is a list of sources at the bottom of the page, but no references throughout the page. It's not at all clear which statements are well-sourced and which are original research.
In particular, the section on the effect as a logical fallacy seems nice, but is it original research or can it be found in one of the sources?
The article now stresses interpretations of computer behavior. I think that's good, but it makes the comments on operator overloading even less coherent. When a user interprets the plus in a computer program, he is not interpreting computer behavior!

Thanks again, Alba, for the hard work on this article. I hope we can fix these remaining issues. Phiwum 18:47, 8 October 2006 (UTC)[reply]

Non-interactive contexts?

Would it be incorrect to say that empathising with a fictional character (outside of interactive fiction) is an instance on the ELIZA effect? You know the character is not real, but you begin to ascribe human motivations to it. So, is the ELIZA effect any different than "the suspension of disbelief"? Somegeek 17:29, 18 October 2006 (UTC)[reply]

Also, if a character is programmed to *model* an emotion or opinion and act accordingly, is it still correct to describe empathy as the ELIZA effect? (The article notes that ELIZA did not simulate emotions such as "interest" since ELIZA doesn't really keep any state around. But it doesn't really address the case of a chatterbot that DOES simulate emotions.) Somegeek 21:33, 26 October 2006 (UTC)[reply]

Operator overloading

I have removed the following paragraph on operator overloading.

The ELIZA effect is also used in the construction of programming languages. The symbol +, for example, is assumed by users to mean 'addition' regardless of context. The symbol + is sometimes also used to represent an algorithm for string concatenation. Program authors can use the + without knowing that it is an overloaded operator that implements two different algorithms. The ELIZA effect can cause users to assume that the program (or programmer) has the same assumptions about the meaning of + that they do.

These comments have nothing at all to do with the ELIZA effect. The ELIZA effect is said to be about misinterpreting computer behavior as intensional. Overloading operators is not about computer behavior, but the meaning of certain symbols in a language. (For that matter, the notion that programmers can use an overloaded operator without knowing it is overloaded is strange indeed! Why would I write "xxx" + "YYY" unless I know that "+" is overloaded to deal with string concatenation?)

If there is some reliable reference discussing the ELIZA effect and overloading, please feel free to cite it and replace this text. Phiwum 12:16, 14 April 2007 (UTC)[reply]

Dispute: The ELIZA effect as logical fallacy.

The article says the following:

The ELIZA effect is a special case of the logical fallacy of affirming the consequent:

If something is motivated by X, it behaves in manner Y.

This program behaves in manner Y.

Therefore, this program is motivated by X.

Even if the program is motivated by X, it does not follow that the observed behavior Y resulted from motivation X. Furthermore, it cannot even be demonstrated that the program is ever motivated by X. Indeed, in many cases, motivation by X is impossible (example: "The program thinks I am attractive".)

Whether the effect is really caused by such thinking is an empirical question. It is just as plausible that the effect is caused by the valid but unsound deduction:

If something behaves in manner Y, then it is motivated by X.

This program behaves in manner Y.

Therefore, this program is motivated by X.

I do not see any evidence that the effect is caused by the former syllogism rather than the latter, which is valid but has a false major premise.

Indeed, we may as well say that all non-solipsists are victims of a fallacy, too, since they apply some reasoning similar to the invalid argument above. It is not enough to notice that one obvious argument for a conclusion is fallacious. We need some evidence that people actually use that fallacious deduction. Phiwum 13:31, 14 April 2007 (UTC)[reply]

Deleted Section

The logical fallacy section has been in the article uncited since October (when I first mentioned that it looks like OR). I think that's long enough. I have deleted the section, which appears below.

  ==Logical fallacy==

  The ELIZA effect is a special case of the logical fallacy of affirming the consequent^{[citation needed]}:

  If something is motivated by X, it behaves in manner Y.
  This program behaves in manner Y.
  Therefore, this program is motivated by X.

  Even if the program is motivated by X, it does not follow that the observed behavior Y resulted from motivation X. Furthermore, it cannot even be demonstrated that the program is ever motivated by X. Indeed, in many cases, motivation by X is impossible (example: "The program thinks I am attractive".)

  The ELIZA effect is a lesser formal fallacy than anthropomorphization, as the computer user knows that the computer is not a human or a complete artificial intelligence. The user nonetheless implicitly assumes the behavior has the same causes as the same behavior would have in a human. The assumption is a fallacy because the computer cannot experience human motives. While the programmer may have had the motivations the user assumes, this cannot be deduced solely from the program's response: the program's behavior may be an unintended side effect.

  The ELIZA effect ends if the user consciously recognizes that the computer cannot be motivated in the assumed manner^{[citation needed]}.

Please forgive my crappy formatting. I would appreciate it if someone with more Wiki skills fixed up the above. Phiwum 12:31, 17 April 2007 (UTC)[reply]

Contradictory claims.

It seems to me that the following two claims are contradictory. Either the apparent contradiction should be explained away or one of the two should be removed.

Upon observation, researchers discovered users subconsciously assuming ELIZA's questions implied interest in the topics discussed, even when they consciously knew that ELIZA did not simulate emotion.
The ELIZA effect ends if the user consciously recognizes that the computer cannot be motivated in the assumed manner.

Neither claim is sourced. Phiwum 13:17, 15 April 2007 (UTC)[reply]

misnomer?

i have spent some hours speaking with various ELIZAs now, and she is the most awful chatterbot i know of. it's strange that this effect is named after her. --Harlequence 01:41, 17 June 2008 (UTC)

The ELIZA/Doctor script was written in the mid-1960s, and the term was coined soon after. Obviously, ELIZA is less complex than later implementations of similar concepts; however, the term of art remains "ELIZA effect". Best, -- 151.191.175.231 (talk) 16:52, 17 June 2008 (UTC)[reply]

Discussion of gender effects

Near the end of this article, there is a discussion of gender effects related to the ELIZA effect. That section as currently written gives the incorrect impression that Weizenbaum observed something gendered in the way that people respond emotionally to chatbots. However, the references in that section do not actually attribute this idea (or any ideas about gender) to Weizenbaum. The section should therefore be edited to make this more clear.

I'm not convinced that gender needs to be discussed at all in this article. Obviously many things about conversations and emotions entail gender, but I don't know if any of those gender entailments are specific to the conversations that people have with chatbots. Sheldon Rampton (talk) 18:55, 13 June 2022 (UTC)[reply]

I just read the paper, and found that part of the article disturbingly misleading and confusing. You posted a year ago, maybe in the next few weeks I'll try to get around to fixing it up a little bit. I would just remove it, but...

You know, I just checked to find out who put in this phrase: "In defining our proximity to digital assistants through their human attributes, chatbots become gendered entities." Clearly, that requires a citation, or it needs to be clearly marked as an expert opinion, not simply stated as a fact. As it turns out, user Kellychen927 was apparently participating in a classroom project:

https://en.wikipedia.org/wiki/Wikipedia:Wiki_Ed/Northeastern_University/Digital_Feminisms_(Spring_2021)

The edits her account have made seem to support the idea that she was being forced to edit wikipedia. It looks like hundreds of edits in the span of 4 months, 0 outside of that timespan. Hooray for school.

https://en.wikipedia.org/w/index.php?title=Special:Contributions/Kellychen927&target=Kellychen927&offset=&limit=500

Whatever - there are sources, and it's contemporary, so if I ever feel like fixing this page I'll surely leave all that "stuff," just make sure it's properly labeled... 37.48.237.26 (talk) 14:04, 6 July 2023 (UTC)[reply]

Broken Reference Link

Reference 4 to Computer Power and Human Reason by Weizenbaum is broken and doesn't go to anything. 2601:5C7:8302:A5F0:64BA:E1DE:C2E8:1EF (talk) 18:39, 19 December 2023 (UTC)[reply]