Talk:Optimality theory

Linguistics: Theoretical Linguistics C‑class High‑importance

	Linguistics portal This article is within the scope of WikiProject Linguistics, a collaborative effort to improve the coverage of linguistics 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.LinguisticsWikipedia:WikiProject LinguisticsTemplate:WikiProject LinguisticsLinguistics articles
C	This article has been rated as C-class on Wikipedia's content assessment scale.
High	This article has been rated as High-importance on the project's importance scale.
	This article is supported by Theoretical Linguistics Task Force.

Archives

June 2004 to October 2008

Capitalization

WP:MOSCAP doesn't seem to cover cases like this, but the Chicago Manual of Style does: "Names of laws, theories, and the like are lowercased, except for proper names attached to them: Avogadro's hypothesis, the big bang theory, Boyle's law, (Einstein's) general theory of relativity, Newton's first law." —An gr 10:12, 14 March 2009 (UTC)[reply]

I see. Well, it looks weird since it is usually capitalized in the ling literature. I guess I would reject Chicago style. But, I'm not overly concerned with punctuation (but apparently enough to bother editing in the 1st place...). – ishwar (speak) 11:03, 14 March 2009 (UTC)[reply]

My preference is to use "OT" everywhere after the first mention, but someone expanded all the abbreviations, claiming doing so made the article more accessible to laymen or something. I don't quite see how, though; if you know the article you're reading is called "Optimality theory" and if the opening sentence says "Optimality theory (also called OT) is...", then you don't have to be a linguistics expert to figure out what "OT" means in the rest of the article. —An gr 11:51, 14 March 2009 (UTC)[reply]

Rule of thumb: 5 full form at the start, abbreviate thereafter, but for a sprinkling here and there. Seeing Optimality Theory several times reinforces the term for newcomers, OT is less cluttering for everyone laterJohndanR (talk) 19:28, 3 September 2010 (UTC)[reply]

Not that it really matters, but in almost all linguistics literature both words are consistently capitalized. I'm not convinced that manuals of style have more authority on the issue than do the sources themselves. If you google the term, Wikipedia is the only place where theory is not capitalized. --N-k (talk) 04:44, 28 July 2011 (UTC)[reply]

comparative fringe?

on a different note.

Why are comparative tableaux "fringey"? They arent the most popular format to be sure, but some people use them. (I personally like them, esp. if you're looking at lots of constraints and I-O mappings at once.) – ishwar (speak) 11:07, 14 March 2009 (UTC)[reply]

Well, it seemed fringey to me since I had never heard of such thing (despite having written a dissertation and a Habilitation using OT and despite having taught OT at two universities) until I saw it here at the Wikipedia article. If I've never heard of something in my area of expertise until I see it at Wikipedia, my gut reaction is that it must be something fringey used only by a handful of people. Have comparative tableaux made it into textbook introductions to OT? —An gr 11:54, 14 March 2009 (UTC)[reply]

Comparative tableaux are very nice, and should be added to every university OT course. Nothing 'fringey' or unfringey about them: just common sense. Who cares who introduced them, Prince, or the publisher's janitor?JohndanR (talk) 19:28, 3 September 2010 (UTC)[reply]

Comparative tableaux are only a little recent. I think they were officially introduced in Prince (2002a), 'Arguing Optimality' (which was deleted a while ago as one of the 'unused' references after the comparative tableau was removed). Comparatives have made it into OT textbooks (McCarthy's 2001 Thematic Guide, at least); I'm not sure about more general textbooks - my impression is that most of them don't go into OT in enough detail for comparatives to be very relevant. McCarthy & Prince don't seem like the fringe of the OT-using linguistics community, so my feeling is that the comparative example should probably go back in the article. Does anyone else have strong thoughts about this? -WmGB (talk) 18:31, 15 March 2009 (UTC)[reply]

I've restored the discussion of the comparative tableaux on the strength of their inclusion in the Thematic Guide, but we need to be careful of appeals to authority. Just because Alan Prince was one of the first propagators of OT doesn't mean everything he writes is gospel. Joseph Greenberg and Theo Vennemann were once respected linguists too until they became crackpots no one listened to anymore. —An gr 21:25, 15 March 2009 (UTC)[reply]

A Thematic Guide to Optimality Theory is not quite a textbook: it's a reference book that could accompany a textbook. But, McCarthy has written an OT textbook (2008 Doing Optimality Theory) in which he largely dispenses with the earlier violation tableau in favor of a mixed violation-comparative tableau. This tableau type includes all of the violation information in the violation tableaux and the winner~loser ranking information of the comparative tableaux. There will, then, be no loss of information. Below is an example.

violation
	input	C1	C2	C3
☞	A	*	**	***
	B	*	***!
	C	**!	**

comparative
input	C1	C2	C3
A ~ B		W	L
A ~ C	W		L

mixed
	input	C1	C2	C3
☞	A	1	2	3
	~ B	₁	₃ W	L
	~ C	₂ W	₂	L

Not only that, McCarthy uses himself. You can see his most recent ROA posting: roa.rutgers.edu/view.php3?id=1453 (although this is harmonic serialism and not "classic" OT, this has nothing to do with his choice of tableau format). You can, of course, find students of Prince and McCarthy using comparative tableaux, but if you look through ROA, you can find others using them as well.

In addition to this, Prince (along with Adrian Brasoveanu) in a series of papers has explored the logical system of OT. There he uses the Ws and Ls to define an Elementary Ranking Condition (ERC) which is the n-tuple of all constraint ranking information provided by a comparison with the optimal candidate and an losing candidate. (E.g. the ERC of optimal A compared with B above is <e, W, L>). He goes on to show that a given ERC with the right properties can logically entail another ERC and gives a mathematical proof of this. (The practical application of this idea is that if an analyst is establishing a ranking argument, then there is no reason to list candidate x in a tableau if x shows less ranking information than a candidate y: the winner~y ERC would entail the winner~x ERC.) – ishwar (speak) 20:20, 16 March 2009 (UTC)[reply]

rule ordering revert

Why did you turn back my contributions on Optimality Theory?

"because regressive assimilation of the final segment always applies first after /ɪ/-insertion, so that it prevents other rules from applying; see also bleeding order"

There 's nothing wrong with this approach, it's just another version of classic order ranking. Greets, Solejheyen (talk) 18:16, 30 May 2009 (UTC)[reply]

So, there is nothing "wrong" with this approach (or probably most approaches). This is not the reason for rejecting it as formulated above. The reason is that there are no such things as rules in OT and there is no derivational process ordering in OT. An output is invalid because of the constraint mechanism, which has ranked constraints (which are not rules either). Your addition misleadingly makes it seem like a derivational ordering of rules is connected to the mechanisms in OT grammars, but it is not since OT lacks both rules and rule ordering.

It may be interesting to readers to compare solutions of classic phonological problems in OT and some version of classic rule-based generative phonology. (This might really be useful since alot of terminology used to describe things are in pre-OT terms even when the theoretical framework is OT, e.g. iterative, right-to-left stress, default-to-opposite stress, counter-bleeding opacity, etc.) But, this really belongs in a different section, not in an explanatory example showing how OT works. This is just confusing, especially to a naive reader who may not realize that OT lacks processes in a derivational sense. – ishwar (speak) 03:08, 31 May 2009 (UTC)[reply]

Hi, I have changed nothing in the text this time, just added a new link. Greets, Solejheyen (talk) 18:02, 31 May 2009 (UTC)[reply]

"not technically a theory" .. "the term 'theory' is used differently here than in physics" - citations needed

Referring to the section on criticisms, within which we read -

Another objection to optimality theory is the claim that it is not technically a theory, in that it does not make falsifiable predictions. The source of this issue is terminology: the term 'theory' is used differently here than in physics, chemistry, and other sciences.

Citations are very much lacking here, and are certainly needed. Nevertheless this is, to my mind, a point worth exploring. Would the original author agree if I compared this with a made-up example assertion about physical science suggestion : "Fourier Analysis is a theory of sound." (It is not. Fourier Analysis is a mathematical framework used in writing a description of a sound but makes no predictions about the nature of sound, or about possible and impossible sounds, nor could ever possibly do so.) Would that be a fair comparison? CecilWard (talk) 21:11, 8 August 2009 (UTC)[reply]

I would imagine that your comparison is basically valid. My impression is that this is a fairly common objection to OT. While I'm not in the position to provide a source I think it's certainly more than the opinion of a single author. Mo-Al (talk) 04:23, 9 August 2009 (UTC)[reply]

I also have the impression it's a fairly common objection to OT, though one I don't understand since OT actually does make falsifiable predictions, certainly to a greater extent than previous approaches to linguistics. Is it just terminology? Is it just because Lexical Phonology and the Minimalist Program don't use the word "theory" in their title that no one sneers, "They're not real theories"? +An gr 07:53, 9 August 2009 (UTC)[reply]

Well arguably the predictions OT makes aren't very strong. I'm still unclear on whether factorial typology is supposed to be exhaustive or not. Besides, since something like ranking all markedness above faithfulness would clearly never exist even though it is a permutation of existing constraints, there must be some assumption that certain permutations are too far-fetched to exist. It may depend on how strongly one buys into the concept of universal constraints. Mo-Al (talk) 16:30, 9 August 2009 (UTC)[reply]

I don't remember very well, but I think I may have been the one who added the sentence in question to the article (if it wasn't me, then I agree with whoever the real author is). My impression is that in linguistics, the word "theory" generally gets used to refer to any set of axioms, whether they meet the criterion of making predictions that are falsifiable or not. Take Feature Geometry, for instance: it doesn't make any falsifiable predictions unless you also define a set of features, and a set of dependency relationships between them, etc. But, I think it would be consistent with standard practice in linguistics to call Feature Geometry a "theory". In that sense, it seems like this is just a different use of the term "theory", and hence really just a question of terminology. That said, I don't know of much in the way of references that criticize OT for not technically being a theory. I know of one manuscript Kochanski 2005, but it's neither peer reviewed nor published anywhere but the author's website, and I'm not sure it's complete (the references section is blank). So, I'm hesitant to list it as a reference in the article. Does anyone know of anywhere else this argument has been seriously made in print? –WmGB (talk) 18:54, 9 August 2009 (UTC)[reply]

Language acquisition

From the current version: "Language acquisition can be described as the process of adjusting the ranking of these constraints."

This makes it sound like language acquisition is ONLY learning the ranking of these constraints, never word learning or other language learning processes. This sentence might be better thus: "Adjusting the ranking of these constraints is one part of language acquisition". Or something. framed0000 03:40, 17 April 2010 (UTC)

Jargon

As a very interested layman, I find this article frustratingly abstruse. Apparently OT is about processes for converting inputs to outputs. But it is never explained when "inputs" or "outputs" are. I assume an "output" is a spoken utterance. I have no idea what an "input" might be in this context. Tesspub (talk) 06:54, 29 August 2010 (UTC)[reply]

It's not bad, as abstruse Wikipedia entries go, especially for linguistics. With the exception of Radford, Hayes and a few others, pedagogy and plain-English explanation are despised disciplines in the field. The standard textbook on Optimality Theory (Kager, Optimality Theory) is in abysmal prose (as are most Dutch [and Flemish Haegeman, Thinking Syntactically], exception Geert's Morphology, textbook-writers/linguists), 11 years old, and so ridden with errors (an opinion reflected also by one of my profs, who himself was not given to gearing down his explanations), that a non-genius in an OT course could not be blamed for asking 'Is this right, wrong, or merely unintelligible?'

But yes, 'not bad' is not necessarily 'good'. You will not find an layman-intelligible explanation of OT anywhere on the internet. —Preceding unsigned comment added by JohndanR (talk • contribs) 19:03, 3 September 2010 (UTC)[reply]

I've added a few sentences to the start of the article to try to clarify that 'input' is usually an underlying form, and 'output' is the surface form it comes out as.WmGB (talk) 05:04, 27 January 2011 (UTC)[reply]

Article categorization as 'Phonology'

I see that this page has been removed from the Category of Linguistics, and I think this is perhaps a misleading designation. OT is in fact used for things other than phonology: morphology and syntax in particular, and I think for semantics & pragmatics too (albeit less widely), and the ROA even had at least one paper applying OT to Catalan translation studies. That's not mentioned in the article as it is now. I'm starting a section on use of OT outside of phonology; can anyone add more on other such applications?WmGB (talk) 15:19, 15 September 2011 (UTC)[reply]