Open main menu

User talk:Dan Polansky

Removing off-topic imagesEdit

In diff, I removed an image of George Orwell from newspeak entry. The image does not help show what newspeak is at the slightest. Not any marginally related image should be used. Similarly, Pythagorean theorem should not have an image of Pythagoras, but it can have File:Pythagoras-proof-anim.svg. --Dan Polansky (talk) 13:54, 3 February 2018 (UTC)

  Support --Daniel Carrero (talk) 14:13, 3 February 2018 (UTC)
  • In diff, I removed an image of theologian Saint Jerome from avoid like the plague entry. The rationale is similar to the one above: the image adds nothing to understanding the phrase, and is only marginally relevant. --Dan Polansky (talk) 17:27, 23 February 2018 (UTC)
    A couple of days ago I was thinking about removing the image from another WOTD page, down but not out. The entry presents idiomatic meaning of the phrase, while the picture shows somebody who seems literally down AND out. What do you think about it? --Jan Kameníček (talk) 17:51, 23 February 2018 (UTC)
    Thank you. In diff, I removed the image from down but not out. However, having done that, I think it is more of a borderline case since at least, the image does show someone who s down but not out. Still, I would argue that the "not out" is not obvious from the image, and that the image, instead of adding clarity, itself needs clarifying. In any case, I find the caption of the image very bad, far removed from the objective of serving the uses cases of a dictionary. --Dan Polansky (talk) 17:59, 23 February 2018 (UTC)
  Support. It might be more useful to engage in actual dialogue with @Sgconlaw about this. —Μετάknowledgediscuss/deeds 21:57, 23 February 2018 (UTC)
Yes, better than doing it behind his back. DP has a "thing" about this subject, I notice. DonnanZ (talk) 22:10, 23 February 2018 (UTC)
Frankly, I do not see what the big problem is, particularly if the image is of someone or something which is linked to the origin of the entry such as St. Jerome or Pythagoras. I also think that readers may be interested in knowing more about an image rather than less. What is the point of, say, having a photograph of a lake with the caption "A lake", rather than something like "Lake Superior is the largest of the Great Lakes in North America"? We all learn something more from a caption along the latter lines than the former. — SGconlaw (talk) 03:45, 24 February 2018 (UTC)
I am irritated to see irrelevant clutter. I am irritated to see an image that is only marginally relevant, an image for an image, as it were, one that gives the impression that it enables someone to check a checkbox in some checklist or something.
I expect images to inform me about the referent or give a better idea of meaning. The point is in the image, not in its caption. Readers should not be informed about images; rather, images should inform the readers about the words. Images are not the thing described in the dictionary; they are the description, or in their case, depiction. The things described are the lexicographic entities, such as word meanings.
The following image caption that was in avoid like the plague irritates me:
"Sv. Hieronim (Saint Jerome, 1725–1735) by Giuseppe Antonio Petrini, from the collection of the National Gallery of Slovenia in Ljubljana, Slovenia. The term avoid like the plague is believed to originate from similar phrases in Latin used by Jerome in his letters."
I don't care about Giuseppe Antonio Petrin or National Gallery of Slovenia; that has nothing to do with the word or phrase. I find the links to 1) Giuseppe Antonio Petrini and 2) National Gallery of Slovenia to Wikipedia entirely inappropriate and a distraction. It looks like some kind of association game: let's make a chain of associations between entities where we quickly veer off to what is a tangent at best. Let's start with any physical object (a chair, a mountain, an apple), and let's show a painting showing that object. That now enables us to talk about things that have nothing to do with the object, like the author of the painting or the photograph showing the object, and the place in which the painting or photograph is shown. We end up with captions that, regardless of the kind of the physical object, talk of things that have nothing to do with the physical object per se but rather are attributes of its representation. I made a mention of captions in March 2017 at User_talk:Sgconlaw#Long_image_captions.
If the reader is interested to know more about the image, in a dictionary, they are in the wrong place since the dictionary informs about words, not images. The reader can learn more about the image by clicking on the image and navigating to Commons, a project hosting images as its core objects, in which detailed information about the image can naturally be hosted. --Dan Polansky (talk) 06:38, 24 February 2018 (UTC)
I agree with Dan Polansky. --Per utramque cavernam (talk) 09:42, 24 February 2018 (UTC)
I sympathise with SGConlaw. He puts a lot of effort into WOTD, and finding an appropriate image for any entry can be quite a challenge, let alone for avoid like the plague, bearing in mind that we are limited to what is available on Wikimedia Commons. Images tend to make an entry more interesting. DonnanZ (talk) 10:57, 24 February 2018 (UTC)
Thank you. — SGconlaw (talk) 02:55, 25 February 2018 (UTC)
Indeed, for many kinds of entries, finding an appropriate image is a challenge, or is impossible. We should not aim at having at least one image in every entry. --Dan Polansky (talk) 11:03, 24 February 2018 (UTC)
  Support Relevant images are very helpful and a way to set us apart from traditional print. Irrelevant ones are clutter. Equinox 23:43, 25 February 2018 (UTC)
Conditional   Support. It seems a lot of the entries with only marginally relevant images are WOTD entries. WOTD itself arguably has marginal relevance to the actual project, as do these images. So, I wonder if one of two things can happen:
  1. The image in question would only stay up on the day that the entry was WOTD.
  2. Perhaps the image can show up in the WOTD form itself on the main page.
I do still find these images entertaining, but it is indeed true that entertainment is not the overall purpose of a dictionary. I guess usefulness is priority 1, while entertainment is priority 2. So, being priority 2, the images should only stay under conditions similar to those mentioned above IMO if at all. PseudoSkull (talk) 02:17, 26 February 2018 (UTC)

I wonder what you would think of the relevance of the image at 𐤉𐤓𐤅𐤔𐤋𐤉𐤌‎. It is not an illustration of the meaning, but of an attestation. I personally think this is a good use of images. --WikiTiki89 19:54, 17 April 2018 (UTC)

@WikiTiki: I think it's fine. It's lexicographical. It's like an attestinq quotation, just that instead of encoded text (ASCII or Unicode), you have an image. --Dan Polansky (talk) 15:50, 20 April 2018 (UTC)
I agree. --Per utramque cavernam (talk) 18:16, 25 April 2018 (UTC)
That kind of thing does seem to make more sense for ancient languages though (where cites in "normal text" might be harder to find). Equinox 18:19, 25 April 2018 (UTC)

Excessive image captionsEdit


I touched on the subject of excessive image captions in #Removing off-topic images, where it did not really belong. Put briefly, readers should not be provided information about images; rather, the images should provide information about word referents. If the reader is interested to know more about the image, in a dictionary, they are in the wrong place, and should click on the image to go to Commons instead.

barbeau shows my idea of a fine caption. It says "barbeau". The caption could also be empty, but it looks a bit odd since there is an empty space intended for the caption. Still, the empty caption could just need some getting used to, and could be fine as well. --Dan Polansky (talk) 09:40, 24 February 2018 (UTC)

Duden:Mops[1] has an image with no caption. What's not to like. Maybe a little frame would make it look neater. --Dan Polansky (talk) 11:24, 24 February 2018 (UTC)

With barbeau there's little more to add. In other cases I think it is helpful to give some explanation. I won't point out any examples for fear of tampering. DonnanZ (talk) 11:30, 24 February 2018 (UTC)
Clicking on an image to find more info doesn't always help. Sometimes an image can be used to illustrate more than one word, if both subjects appear in the same image. DonnanZ (talk) 11:49, 24 February 2018 (UTC)
That's an argument to use a caption that disambiguates when that is required, not an argument to have a caption that gives you the author and the gallery of a painting showing the referent. --Dan Polansky (talk) 11:54, 24 February 2018 (UTC)
I added now a third image of barbeau without any caption, and also without any space intended for a caption. Does it look good? --Daniel Carrero (talk) 16:30, 3 March 2018 (UTC)
Thanks. That looks excellent, especially since the image has a white background. It may look a little worse for someone who tweaks their CSS to have a dark background. --Dan Polansky (talk) 16:38, 3 March 2018 (UTC)
You're welcome. FWIW, I support adding frameless images like that 3rd barbeau in entries when applicable. --Daniel Carrero (talk) 16:49, 3 March 2018 (UTC)

There is now Wiktionary:Beer_parlour/2018/April#Image captions and descriptions of the representing objects such as paintings. --Dan Polansky (talk) 08:53, 1 April 2018 (UTC)


Hi, just to explain my attitude to the gloss template. Compare e. g. the Czech entry on Ceres without and with the template:

  1. (Roman mythology, feminine) Ceres (Roman goddess of agriculture; equivalent to the Greek goddess Demeter.)

Here you can see that the most important thing, the link to the English entry, is drowned between two pairs of brackets. In my opinion the following solution looks better:

  1. (Roman mythology, feminine) Ceres, Roman goddess of agriculture; equivalent to the Greek goddess Demeter.

I do not think it is confusing or less comprehensive to the reader. --Jan Kameníček (talk) 09:25, 4 March 2018 (UTC)

The current conventional format in Czech entries is easier for the eye to scan, IMHO:
  1. (Roman mythology, feminine) Ceres (Roman goddess of agriculture; equivalent to the Greek goddess Demeter)
It typographically separates the line into two parts: 1) a list of candidate translations, 2) a disambiguating gloss. I find it preferrable, but it is probably a matter of taste. The gloss does not contain links. The format invites the reader to click on "Ceres" and nowhere else, which I find to be a benefit; it is in Ceres where the most relevant other information can be found, including other senses that can confound the reader of a translation that uses "Ceres".
It is questionable whether "(Roman mythology, feminine)" should be there at all, but that is a slightly different discussion. There should ideally be no links in "(Roman mythology, feminine)", but someone changed it a long time ago without a discussion AFAIK, and I do not have the energy to fight it.
That said, by my lights, you have the right to use the format that you prefer, in the new entries that you create.
--Dan Polansky (talk) 09:35, 4 March 2018 (UTC)


Lojban is a constructed language whose development began in 1987 and the first "baseline" was completed in 1997, per WP. Its iso code is jbo. It has peculiar parts of speech: brivla, cmavo, gismu, lujvo, and rafsi, reflected in Wiktionary categories such as Category:Lojban brivla. There is Wiktionary:About Lojban. Current WT:CFI requires 3 independent quotations in use for attestation since Lojban is listed in Wiktionary:Criteria for inclusion/Well documented languages. Category:Lojban lemmas has 2,185 entries. Wikipedia written in Lojban has 1,207 articles.


Some links:

--Dan Polansky (talk) 11:02, 10 March 2018 (UTC)

Copyright and lexicographical dataEdit

Meta:Wikilegal/Lexicographical Data, created on 17. 2. 2018, is a pretty interesting write-up on the subject. In particular, "There can be a copyright on the definitions of words as long as they are creative", but also, "Most lexicographical data is factual in nature and, therefore, cannot be copyrighted as a standalone work." --Dan Polansky (talk) 20:05, 10 March 2018 (UTC)

Inclusion of theoremsEdit

Tauberian theorem is now at WT:RFDE, which will be archived at Talk:Tauberian theorem.

We have Pythagorean theorem. Pythagorean theorem is not a sum of parts; you cannot know which theorem it is just from "Pythagorean" and "theorem". Pythagorean theorem is in Merriam-Webster[2], AHD[3] and[4].

Category:English eponyms contains many similar items of the form (person-name-adj) (class name), including Abelian algebra.

A counterargument could be like this: You cannot know what food items are part of "Chinese cuisine" just from "Chinese" and "cuisine", and therefore, following an argument similar in form, you would have to include Chinese cuisine. My response would be that "Pythagorean theorem" is a result of naming or christening whereas "Chinese cuisine" is not.

A related discussion is at Talk:Pauli exclusion principle, where I posted "Other entries for laws, principles, conjectures and theorems: Bragg's law, Kepler's laws, Metcalfe's law [...]".

--Dan Polansky (talk) 19:50, 16 March 2018 (UTC)

Czech syllable counterEdit

Here is my Czech syllable counter that I use for rhyme pages. It is a heuristic, and it sometimes determines the wrong number of syllables. I feel it is better than nothing; doing it manually would be horribly slow. It is written in Excel VBA. Feel free to reuse; all talk page content is obviously licensed by the same license as mainspace: CC-BY-SA.

Private Function SyllableCount(vString As String)
  Set ForeignKeySheet = Sheets("SyllableCount")
  Set FoundCell = ForeignKeySheet.Columns(1).Find(vString, LookAt:=xlWhole)
  If Not FoundCell Is Nothing Then
    SyllableCount = FoundCell.Offset(0, 1)
    Exit Function
  End If
  SyllableCount = NumberOfVowels(vString)
End Function

Private Function NumberOfVowels(vString As String)

  Dim LCaseString As String
  Dim Char As String
  Dim CharBefore As String
  Dim CharAfter As String
  'Determine the first approximation by counting a, e, i, o, etc.,
  'and by counting r or l between two consonants
  Result = 0
  LCaseString = LCase(vString)
  For Idx = 1 To Len(LCaseString)
    Char = Mid(LCaseString, Idx, 1)
    If IsVowel(Char) Then
      Result = Result + 1
    ElseIf Idx > 1 And Idx < Len(LCaseString) And _
      (Char = "r" Or Char = "l") Then
      If Not IsVowel(Mid(LCaseString, Idx - 1, 1)) And _
         Not IsVowel(Mid(LCaseString, Idx + 1, 1)) Then
        'r or l between two consonants, as in tržnice or mlžit
        Result = Result + 1
      End If
    End If
  ' Adjust the number by diphthongs (au, eu, ou): treat sometimes as one vowel.
  ' Ou
  OuPosition = InStr(LCaseString, "ou")
  If OuPosition = 1 Or OuPosition = (Len(LCaseString) - 1) Then
    Result = Result - 1 'Treat ou as one syllable; assume only one "ou" occurence, which is wrong
    ' dlouhou - counterexample to one "ou"
  ElseIf OuPosition > 0 Then
    CharBefore = Mid(LCaseString, OuPosition - 1, 1)
    CharAfter = Mid(LCaseString, OuPosition + 2, 1)
    If CharBefore = "d" And _
       CharAfter = "č" Then
      'Leave the result alone; example: doučit
    ElseIf CharBefore = "p" And _
       (CharAfter = "č" Or CharAfter = "ž" Or CharAfter = "k" Or _
        Mid(LCaseString, OuPosition + 2, 2) = "sm") Then
      'Leave the result alone; examples: poučit, použít, poukaz, pousmát (vs poustevník)
      Result = Result - 1
    End If
  End If
  AuPosition = InStr(LCaseString, "au")
  If AuPosition = 1 Or AuPosition = (Len(LCaseString) - 1) Then
    Result = Result - 1 'Treat au as one syllable; assume only one "au" occurence, which is wrong
  ElseIf AuPosition > 0 Then
    CharBefore = Mid(LCaseString, AuPosition - 1, 1)
    CharAfter = Mid(LCaseString, AuPosition + 2, 1)
    If CharBefore = "n" And _
       (CharAfter = "č" Or CharAfter = "k") Then
      'Leave the result alone; examples: naučit, nauka
    ElseIf CharBefore = "z" And _
       (CharAfter = "č" Or CharAfter = "v") Then
      'Leave the result alone; examples: zaučit, zaučovat, zauvažovat
      'Reason for char after constraint: zautomatizovat
      Result = Result - 1
    End If
  End If
  'No adjustment. E.g. pneumatika seems to have 4 syllables, and neuróza 4 syllables.
  NumberOfVowels = Result 'Return
End Function

Private Function IsVowel(vString As String) As Boolean
  char = vString
  If char = "a" Or char = "e" Or char = "i" Or _
     char = "o" Or char = "u" Or char = "y" Or _
     char = "á" Or char = "é" Or char = "ě" Or _
     char = "í" Or char = "ó" Or char = "ú" Or _
     char = "ů" Or char = "ý" Then
    IsVowel = True
    IsVowel = False
  End If
End Function

--Dan Polansky (talk) 17:21, 17 March 2018 (UTC)

As is apparent from the above code, what causes trouble are the diphthongs (au, eu, ou). The number of syllables is the number of vowels or things that act as vowels. Things acting as vowels include certain occurrences of r and l; that seems to be handled well. Disphthongs also act as vowels, but only when they are really diphthongs. Like, we have nauka vs. mouka; nauka is na-uka, 3 syllables, while mouka has 2 syllables. For diphthongs, the script has to resort to unreliable heuristics. Words containing merely apparent diphthongs include doučit (vs. doufat, doutnat, doupě, doušek), nauka, naučit (vs. Nautilus), poučit, poupravit, použít, poukaz, poukázat, pousmát (vs. pouhý, poupě, pouze), zaučit, zaučovat, zauvažovat. --Dan Polansky (talk) 09:28, 25 March 2018 (UTC)

Code updated, especially for treatment of au after z. --Dan Polansky (talk) 10:26, 25 March 2018 (UTC)

Happy EasterEdit

I wish to you and all users of Wiktionary Happy Easter.

Leonard Joseph Raymond (talk) 20:49, 1 April 2018 (UTC)

Outstanding policy issuesEdit

Some outstanding policy issues:

  • Promote Inflection heading and deprecate Declension and Conjugation heading; support very uncertain.
  • Synonym content to Thesaurus: Find out whether people want to move as much content as possible away from mainspace to Thesaurus.
  • Entry name conventions for non-English Thesaurus entries: find agreement.
  • Add section on Images to CFI, at first only stating what we already do and what is uncontroversial.
  • Find out whether we could unify capitalization of English definitions on lowercase without final period or on capitalized first letter and final period. Lowercase without final period is online in M-W, Collins, and Macmillan. Uppercase with final period is online in Webster 1913, AHD, and Seems very arbitary, a matter of taste.
  • Allow Noun phrase and Verb phrase part-of-speech headings for highly inflected languages; support very uncertain; desirability also uncertain. Rationale: In highly inflected languages, part of speech is not only about the syntactic role of a term in a sentence but also about its inflection, and a noun phrase does not inflect like a noun.

--Dan Polansky (talk) 16:06, 7 April 2018 (UTC)

sound postEdit

Hello, do you think you could please add 'sound post' as a meaning of 'duše' if you do agree that it's correct. I would really appreciate it. Thanks! Kaixinguo~enwiktionary (talk) 18:09, 7 April 2018 (UTC)

I could, but I have no source telling me "sound post" can be translated as "duše", or that "duše" sometimes means as much. What is your source? --Dan Polansky (talk) 18:12, 7 April 2018 (UTC)
I found it in SSJC, sense 6, and so I expanded duše entry. I hope they did their attestation homework. --Dan Polansky (talk) 18:16, 7 April 2018 (UTC)
Oh, I'm sorry, I thought you would know it but it's true that these words are a bit specialised. I don't have any source at all except the page 'sound post' here. If you feel it's potentially inaccurate then I will remove it; I can check with a Czech musician or violin maker. Kaixinguo~enwiktionary (talk) 18:21, 7 April 2018 (UTC)
No, it's fine. The meaning is in SSJC[5]. They could have made a mistake, of course, but that would be a rare event. --Dan Polansky (talk) 18:23, 7 April 2018 (UTC)
Thank you. :D Kaixinguo~enwiktionary (talk) 18:28, 7 April 2018 (UTC)

older sisterEdit

I somewhat agree with this edit; the wording of {{&lit}} is confusing when there's only one definition. However, defining "older sister" by "A sibling's older sister" seems circular to me. --Per utramque cavernam (talk) 18:59, 17 April 2018 (UTC)

I think my version "(translation hub) A sibling's older sister" is perfect (no longer there). The only thing missing is a link from "translation hub" to a glossary where the idea of translation hubs is explained, probably Appendix:Glossary. The definition is either circular or trivial, sure, but that is what you expect from a translation hub. --Dan Polansky (talk) 15:53, 20 April 2018 (UTC)

Czech related terms derived from LatinEdit

What follows is a design for section Related terms for Czech words derived from Latin:

  • Let us identify related term (RT) rings. Let these be grouped by a base Latin word having no apparent prefix, e.g. codex.
  • Let a RT ring include terms differing in suffix but also terms differing in prefix. Thus, for kód, let us include not only kodex in the RT ring but also dekódovat and zakódovat.
  • Let a RT ring also contain terms originating as compounds. Thus, let the ring for reálný contain republika = res + public. As a consequence, RT rings may overlap: a term can be member of two RT rings.
  • For each RT ring, let us pick a short Czech term to host the ring, a term that is a member of that ring. Let us place the RT ring into that word's entry. Picking the host may turn challenging in some cases, but let us see. Let the choice of the term be guided by the number of syllables, then number of letters, but let an override be granted.
  • Let other members of the ring point to the host via "===Related terms===(par)* See (host)". This eliminates duplication.

I created reálný and kód as central hosts of RT rings. For reálný, I originally considered reál for its being shorter. However, reál has two etymologies, which makes it a little less suitable as the central host.

The above is described for descendants from Latin, but works for those from Ancient Greek as well.

An alternative would be to create categories, like "Czech terms descended from Latin codex". Then, each member of the ring could have RT display populated from the category. On the downside, this would require that all ring members have entries already created.

Appendix:English words by Latin antecedents will be useful for RT ring identification.

For English, it would work similarly. For example, we start with Latin ambulo. We identify the RT ring, which would be, perhaps incomplete, amble, ambulance, ambulant, ambulate, ambulator, ambulatory, funambulist, perambulate, perambulation, perambulator, preamble, somnambulist, somnambulate, and somnambulistic. We note that amble is the word with the smallest number of syllables, and pick it to host the RT ring. We place the RT ring to amble and let other members of the ring point to amble.

--Dan Polansky (talk) 09:37, 21 April 2018 (UTC)

Using the above design, I created RT rings for akt (ago), verš (verto), dynamo (δύναμαι), takt (tango), mód (modus), lekce (lego), pres (premo), sex (sexus) and póza (pono). When creating póza, I discovered fr:poser has a similar RT ring, in its Apparentés étymologiques: apposer, apposition, déposer, déposition, exposer, exposition, imposer, imposition, impôt, pose, posé, posément, poseur, position, repos, reposer, supposer, supposition. --Dan Polansky (talk) 17:30, 24 April 2018 (UTC)

I created mise (mitto) even though mše has fewer syllables. mise seems to better fit to komise, demise, etc. in its outward appearance and two syllables are not too many; mise has no prefix. --Dan Polansky (talk) 17:17, 25 April 2018 (UTC)

I usually go back to Latin or Ancient Greek base words and stop there. But one can go deeper, to PIE, obtaining a category like Category:English terms derived from the PIE root *steh₂-, which contains both state (from La sto) and static (from ἵστημι). When we would go deeper, we would get Latin, Ancient Greek and Slavic roots (terms that gave birth to descendants) united under one umbrella, it seems; Czech verb stát would end up in the same group as Czech statický. This might lead to too large RT rings, but I don't know. The rings would disregard AGr vs. La vs. Sl origin. --Dan Polansky (talk) 12:34, 1 May 2018 (UTC)

For Latin iacio, I chose objekt as the host of the RT ring. I don't like it since it contains a prefix, but I found no better candidate; there is no *jekt or *jekce in Czech, from what I found. In the RT ring, objekt has two syllables, and among those that have two syllables, it is first in alphabet. At the end of the day, it would be acceptable, while not nice, if the host were be chosen wholly arbitrarily; the basic job of hosting a RT ring so that other entries can point to it can be done by any entry. --Dan Polansky (talk) 08:23, 6 May 2018 (UTC)

For Latin sisto, I chose existence. It was rather arbitary; I felt existence is more important than asistence. --Dan Polansky (talk) 16:42, 6 May 2018 (UTC)

For Latin volvo, I could have chosen volt or volta, both shorter than voluta, which I ultimately chose. All three are prefixless, so far so good. What speaks for voluta is the "volu" part, present in multiple members of the ring: devoluce, evoluce, involuce, revoluce, etc. --Dan Polansky (talk) 06:51, 8 May 2018 (UTC)

For Latin caput, kápo is shorter than kapitán and kapitál is earlier in alphabet. And yet, I chose kapitán; kápo has a meaning relating to nasty things, and kapitán seems in a better general esteem than kapitál. As for morphology, both kapitán and kapitál contain kapit-, common to multiple members of the ring. --Dan Polansky (talk) 10:50, 8 May 2018 (UTC)

For Latin fero, I included the relace set as part of the RT ring in oferta. This was based on the notion that relatus is from refero. But some sources say that relatus is used as a past participle of refero, which could mean that, morphologically speaking, it is not really stemming from it. If so, relace would be the base of a separate RT ring. I am leaving the oferta ring for fero as is, including relace, relativní, etc., but the question may be revisited later. --Dan Polansky (talk) 14:52, 8 May 2018 (UTC)

Appearance of Ancient Greek reference templatesEdit

I see the following appearance in δύναμαι, except that I do not show hyperlinks:

  • δύναμαι in Liddell & Scott (1940) A Greek–English Lexicon, Oxford: Clarendon Press
  • δύναμαι in Liddell & Scott (1889) An Intermediate Greek–English Lexicon, New York: Harper & Brothers
  • δύναμαι in Autenrieth, Georg (1891) A Homeric Dictionary for Schools and Colleges, New York: Harper and Brothers
  • δύναμαι in Bailly, Anatole (1935) Le Grand Bailly: Dictionnaire grec-français, Paris: Hachette
  • δύναμαι in Cunliffe, Richard J. (1924) A Lexicon of the Homeric Dialect: Expanded Edition, Norman: University of Oklahoma Press, published 1963
  • δύναμαι in the Diccionario Griego–Español en línea (2006–2018)
  • δύναμαι in Slater, William J. (1969) Lexicon to Pindar, Berlin: Walter de Gruyter
  • G1410 in Strong’s Exhaustive Concordance to the Bible

Related: User talk:Dan Polansky/2017#External link templates and excessive detail.

--Dan Polansky (talk) 08:22, 28 April 2018 (UTC)

I removed copyright sign from Diccionario Griego–Español en línea: we don't do that in general, and that is not a usual referencing practice. --Dan Polansky (talk) 08:35, 28 April 2018 (UTC)

I see the following appearance of English reference templates in climate, except that I do not show hyperlinks:

  • climate in Webster’s Revised Unabridged Dictionary, G. & C. Merriam, 1913
  • climate in The Century Dictionary, The Century Co., New York, 1911
  • climate at OneLook Dictionary Search

--Dan Polansky (talk) 13:18, 6 May 2018 (UTC)

I see the following appearance of Latin reference templates in emo, except that I do not show hyperlinks:

  • emo in Charlton T. Lewis and Charles Short (1879) A Latin Dictionary, Oxford: Clarendon Press
  • emo in Charlton T. Lewis (1891) An Elementary Latin Dictionary, New York: Harper & Brothers
  • emo in Gaffiot, Félix (1934) Dictionnaire Illustré Latin-Français, Hachette
  • Carl Meissner; Henry William Auden (1894) Latin Phrase-Book‎[1], London: Macmillan and Co.

--Dan Polansky (talk) 13:20, 6 May 2018 (UTC)

Above we can see multiple reference templates that show a non-English title with no English rendering: Bailly 1935, Diccionario Griego–Español en línea, and Dictionnaire Illustré Latin-Français. I think it is fine for the languages involved, which are Romance languages; it could be less nice for e.g. Russian. Since a lot of English vocabulary stemms from Latin, the meanings of the titles from Romance languages are fairly easy to guess for a native English speaker. --Dan Polansky (talk) 10:15, 8 May 2018 (UTC)

Another case of no English title is this:

  • labor in Ramminger, Johann (accessed 16 July 2016) Neulateinische Wortliste: Ein Wörterbuch des Lateinischen von Petrarca bis 1700‎[2], pre-publication website, 2005-2016

It is fine with me (except for the "accessed" thing) since I know German; for me, even romanized Russian is fine. I admit that it is not so user friendly to those who speak no German. A link to an appendix providing detail would handle the need. Adding an English rendering for the mainspace would make it rather long. Anyway, here's rendering of that title by Google Translate: "Neo-Latin word list: A dictionary of Latin from Petrarca to 1700". Not too shabby; Neo-Latin is New Latin, of course. --Dan Polansky (talk) 12:25, 8 May 2018 (UTC)

I see the following appearance of English reference templates in abridgement, except that I do not show hyperlinks:

  • abridgment in Webster’s Revised Unabridged Dictionary, G. & C. Merriam, 1913
  • abridgment in The Century Dictionary, The Century Co., New York, 1911
  • abridgment at OneLook Dictionary Search

--Dan Polansky (talk) 18:16, 8 May 2018 (UTC)

Attributive nounsEdit

My position is that attributive positions are adjectival positions. An English item is attested as a noun if it is attested (1) in the subject position, or (2) with an indefinite article, or (3) in plural; there may be more clear tests. lumbar is an example of an adjective that seems attested predominantly or only in attributive position and is non-comparable. It seems to follow that the position or doctrine that a non-comparable item only attested in an attributive position is not shown to be an adjective is untenable. There is no ultimate adjectivity test for English, as far as I know; there are only partial tests that decide some cases one way or another, like: if it is comparable, it is an adjective.

M-W regards attributive nouns as nouns: Nouns That Look Like Adjectives, It does admit they do some of the work adjectives do: "Attributive nouns do some of the same work that adjectives do, but that doesn't mean they're not nouns. Think of them as nouns that learned how to multitask." The source was pointed out by Per utramque cavernam; thanks.

Another M-W article is Attributive nouns in Help, One quote is this: "While any noun may occasionally be used attributively, the label often attributive is limited to those having broad attributive use."

Another source is ATTRIBUTIVE NOUN, Concise Oxford Companion to the English Language, originally published by Oxford University Press 1998, It says "ATTRIBUTIVE NOUN. A noun that modifies another noun: steel in steel bridge; London in London house. Nouns used in this way are sometimes said to be adjectives or to behave like adjectives".

--Dan Polansky (talk) 17:37, 1 May 2018 (UTC)

When a multi-word noun phrase is used attributively, it is often hyphenated. For hyphenated forms of nouns, e.g. geographical-area, we have {{attributive form of}}. These seem to be mostly ranked as nouns in our mainspace. The template is used not only in hyphenated form; it is used e.g. in pyjama. --Dan Polansky (talk) 19:04, 1 May 2018 (UTC)

Quick note (feel free to (re)move): The binary adjective is still, in a mathematical sense, comparable. Even if no comparative or superlativ form exist, the negation should still qualify as comparison, if equality is defined at least. So, "comparable" can be contrasted with "incomparable", "true" with "false", etc. The same works with nouns and verbs, so I am not sure this helps. Rhyminreason (talk) 20:48, 19 May 2018 (UTC)

Later: Deletion discussions for attributive nouns include Talk:transitive-verb, and Talk:cookie-cutter; a RFDE discussion "All English attributive forms (with hyphens) of noun phrases" is archived at Talk:transitive-verb# as well. --Dan Polansky (talk) 18:43, 1 July 2018 (UTC)


Hey. Can you make an entry for vít please? It's linked from výt --Cien pies 6 (talk) 13:34, 2 May 2018 (UTC)

Not so easy. One source tells me that it means "to make a wreath", so the translation could be wreath (verb), but I am not sure. I imagine that if you have a stick and a wire, then the operation by which you turn the wire around the stick to create a coil would be "vít", but here again I am not sure enough. --Dan Polansky (talk) 18:04, 4 May 2018 (UTC)
An aside: There is a test item for Czech learners, especially children, which requires them to write Víly vily věnce, given only the auditory. The hard part is the use of i vs. y. --Dan Polansky (talk) 18:08, 4 May 2018 (UTC)


Here is a warning that I am more accustomed to giving newcomers, but apparently you need to hear it: do not create entries in languages you do not know and have not studied. —Μετάknowledgediscuss/deeds 17:07, 14 May 2018 (UTC)

I have a long-time practice of creating entries in certain languages, including Latin and Ancient Greek. I am very careful with sources. Years ago, I even created User:Dan Polansky/Verification so that those who have doubts can verify my work. Except for one report at the beginning which turned out to be of questionable validity, I have not received any reports of errors that I have made. Naturally, I will yield to a verifiable consensus on the matter. --Dan Polansky (talk) 08:07, 15 May 2018 (UTC)
You missed the fact that your gloss for the adjective κλινικός (klinikós) was in the wrong part of speech (you wrote the gloss as a noun when the word wasn't a noun) for no good reason, and when that was corrected, you reverted it back to the wrong part of speech. mellohi! (僕の乖離) 15:54, 15 May 2018 (UTC)
I entered bed as the definition since that is used as an adjective as well. Someone may like it more explicit, but again, English bed acts not only as a noun but also as an adjective. And it was not "corrected"; it was replaced by Metaknowledge with RFDEF; and at the same time, Metaknowledge removed further reading that was relevant to etymology: OneLook. --Dan Polansky (talk) 16:29, 15 May 2018 (UTC)
That's a noun used attributively, not an adjective. Either way, "of or pertaining to a bed" is more precise. —AryamanA (मुझसे बात करेंयोगदान) 21:11, 15 May 2018 (UTC)
κλινικός is likely to be used predominantly attributively. Attributive uses of nouns are adjectival uses, I maintain, not for the first time. That said, I am not preventing anyone from being "more precise", as they see fit. I merely maintain that what I entered was accurate, and that I have a track record of creating accurate entries in Ancient Greek and Latin based on careful work with sources, and knowledge of lemmatization practices of the English Wiktionary. Careful work with sources is the hallmark of good wiki work, not being a certified expert in an area of contribution. --Dan Polansky (talk) 21:21, 15 May 2018 (UTC)
Careful work with sources is good, but it's not a substitute for studying and having knowledge of the language at hand. κλινικός is actually an adjective (it's declined in all three genders), not an attributive noun. Furthermore, Mellohi added a new noun sense, that seems to not be used attributively. —AryamanA (मुझसे बात करेंयोगदान) 00:04, 16 May 2018 (UTC)
There is no disagreement on whether κλινικός is an adjective; indeed, I entered the word with the part of speech of adjective in the level-3 heading. The disagreement is about English nouns and whether they can be used to define of-or-pertaining-to adjectives. That disagreement has nothing to do with Ancient Greek per se. English adjectives do not decline at all, and therefore they do not decline in gender; nothing interesting going on here. A recent writeup is at #Attributive nouns, containing links to two sources.
As for "studying and having knowledge of the language at hand", it depends on what sort of knowledge is meant. I believe to have enough knowledge of Ancient Greek to create accurate entries, as borne out by my track record. Similarly, I have enough knowledge of Danish to create accurate entries as long as I carefully peruse sources. The "do not create entries in languages you do not know and have not studied" is not a Wiktionary policy, as far as I know, and I believe there is enough evidence to show it is not a common practice either, and I believe there is no consensus to support such a policy. If such a policy were adopted, it would drastically reduce editor ability to create English Wiktionary content based on careful work with sources. --Dan Polansky (talk) 05:21, 16 May 2018 (UTC)
I think this entry was exceptionally bad; you disagree. However, editors who work on Ancient Greek seem to agree with me. So let me rephrase my warning to be clearer: do not create Ancient Greek entries. —Μετάknowledgediscuss/deeds 06:02, 16 May 2018 (UTC)
──────────────────────────────────────────────────────────────────────────────────────────────────── I will yield to consensus. I will also yield to 60%-majority. I am considering yielding to 50%-majority, but I am not sure. That should be based on policy, not on "Dan Polansky must not edit Ancient Greek entries". The entry was accurate and useful: it contained an accurate etymology, accurate yet incomplete semantics, and two good links to external sources so the reader could get a lot of other detail one or two clicks away. One of the links, OneLook, was removed by Metaknowledge two times without an explanation. --Dan Polansky (talk) 06:09, 16 May 2018 (UTC)
You will also, like all other editors, yield to admins. I have left the OneLook link since you seem so ardent about keeping it in, but it is generally considered inappropriate to use a reference for a different term in a different language when superior references exist. —Μετάknowledgediscuss/deeds 06:14, 16 May 2018 (UTC)
What is the online superior further reading as far as etymology? --Dan Polansky (talk) 06:18, 16 May 2018 (UTC)
L&S's Intermediate Greek-English Lexicon makes it explicit, I suppose. —Μετάknowledgediscuss/deeds 06:33, 16 May 2018 (UTC)
Where can I see the κλινικός entry in that lexicon? Is it really as good as for etymology as etymonline[6]? Let us also note that OneLook contains multiple sources providing etymology, which provides independence, including M-W. --Dan Polansky (talk) 06:37, 16 May 2018 (UTC)
To wit: when creating κοσμέω, I consulted three sources for the etymology as indicated in the creation summary, thanks to OneLook. Our reader does not need to know to look in OneLook; I shared with the reader the good-source knowledge that I have. Consulting multiple sources is important since they sometimes disagree on etymology, as I have recently noted on Talk:nota. --Dan Polansky (talk) 06:45, 16 May 2018 (UTC)
I guess the entry is here, and the source is from 1889; OCLC 15599938 is another edition or republishing from 1900. As for etymology, the entry says "κλινικός 1 κλίνη". --Dan Polansky (talk) 07:07, 16 May 2018 (UTC)
For the record, and mostly so that I do not forget: Here is User talk:Razorflame#Languages you don't know, and here is Metaknowledge post there: "Razorflame, everything may be correct, but you may still be violating Wiktionary and/or Wikimedia policy. As an example, some of your Ido entries may be in contravention of WT:CFI; see WT:RFV#alotropa as an example." Subsequently, Razorflame was not blocked for creating entries in multiple languages they did not know, in which they continued; rather, I was blocked for repeated pointing out Razorflame's shoddy editing. The problem was not with the languages, but rather with the long track record of shoddy editing; that is why I made the request to Razorflame. Metaknowledge was involved in my block (not doing the block themselves), expressly approving it. Seems ironic, doesn't it? --Dan Polansky (talk) 07:46, 16 May 2018 (UTC)
Perusing his talk page, it seems that I gave Razorflame a warning, after which he ceased to create any such entries and has been inactive ever since, so there was no need to block him. Looking at your block log, I can't tell which incident you're referring to. My only involvement in your block log has been to reduce the length of one of your blocks and I think to give you talk-page access; I have never actually given you a block and I would be content to keep it that way. —Μετάknowledgediscuss/deeds 19:19, 16 May 2018 (UTC)
The Razorflame talk page shows clearly that what Metaknowledge repeatedly raised was attestation, whereas what I raised was languages Razorflame did not know, which, again, I did because of high error rate. And it is clear that, in the thread I linked, no one said they agreed with me and that Razorflame should stop; instead, Metaknowledge said "Razorflame, everything may be correct ...", as quoted. Let the reader make their own judgment and read it if wished; the evidence is there and is clear. The Razorflame case is most instructive since not only did Metaknowledge not support the policy that he now implies exists or he now implies he can enforce against selected editors even when it does not exist, but also because had such a policy existed, Razorflame could have been dealt with swiftly instead of the protracted drama that lasted several years and involved multiple editors. Here I can say, yes, the putative policy would have been very helpful against Razorflame and similar editors. On the other hand, it would penalize careful editors for mistakes made by careless editors. It would be very much against the wiki spirit, in which expertise does not matter, merely enough care when working with sources, and willingness to learn from one's mistakes. Furthermore, the putative policy is all too likely to lead to considerable proficiency overreporting, for which especially certain cultures are noted. --Dan Polansky (talk) 06:35, 17 May 2018 (UTC)
(outdent) In any case, I feel I am wasting my time and energy. This is not anything a transparent civilized process could not handle: Wiktionary:Votes/pl-2018-05/Proficiency as a prerequisite for contribution. --Dan Polansky (talk) 07:02, 17 May 2018 (UTC)
Just FYI ... the issue of how to handle attributive adjectives of the "of or pertaining to X" type is not obvious. In Russian entries we normally would write them like this (as in посте́льный (postélʹnyj), which has the same meaning as κλινικός (klinikós)):
  1. (attributive) bed
This is because they are most naturally translated as nouns in English that are placed in attribution to another noun, as in "bed rest", "bed sore/bedsore", etc. (IMO, phrases like "bed rest" and "wagon wheel" contain two nouns each and no adjectives.) The "attributive" label links to an appendix that describes what "attributive" means in this context and hopefully clarifies why a noun is being used to gloss an adjective. I agree that just using the noun alone as a definition looks wrong. But how to handle these sorts of adjectives is an issue that's specific to the particular language community at wiktionary. I think the larger issue here is not so much how knowledgeable Dan is w.r.t. Ancient Greek but rather how willing he is to learn and follow the conventions of the Ancient Greek community here at wiktionary. Benwing2 (talk) 15:12, 19 May 2018 (UTC)
Thank you. That's the point: it is the most natural translation. For a long time, I have used the practice of using nouns in definitions of certain Czech adjectives. I used to place "(when used attributively)" behind the noun, but then I figured the reader is not stupid and stopped using even that. I do not insist that this is a practice that Ancient Greek entries must follow, merely that my Ancient Greek entry was accurate and that I did not prevent anyone from changing it to what they find preferable. --Dan Polansky (talk) 06:39, 20 May 2018 (UTC)
The verbal-noun forming -ος is a weak rhyme on the adjectival noun forming -ικός. That's one more reason to gloss the lemma like a noun. The accent on -ικός makes all the difference perhaps, but not by much. In case of doubt, though, it's better to lean on the save side and be as explicit as was advised.
I would mostly chalk up meta's negative tone to tongue-in-cheek referring back to old arguments. Rhyminreason (talk) 16:18, 24 May 2018 (UTC) (redacted).
Later: A somewhat related discussion is at User talk:BD2412#bër, where Metaknowledge deleted bëër created by me to make room for the erroneous bëer created by them; so much for experts. --Dan Polansky (talk) 09:00, 3 June 2018 (UTC)


I was blocked by Wyang for "Disruptive edits". I request an unblock. The alleged disruptive edits are probably my addition to Wiktionary:Votes/pl-2018-05/Proficiency as a prerequisite for contribution of a section "Oppose having this vote". I think adding this section was okay; people need to have a recourse agaist votes that they think should not be running. At the same time, it is inappropriate for a vocal minority to prevent votes from running. The vote, deleted by Wyang for the second time, needs to be undeleted. Wiktionary:Votes/2018-03/Showing romanizations in italics by default is an example of a vote that Wyang did not want running, but the overwhelming supermajority of 77.8% did want to see running. --Dan Polansky (talk) 10:35, 19 May 2018 (UTC)

I don't have much context into this issue but Dan hasn't been blocked for anything for 3 years, so 1 week feels too long. I changed it to 1 day. Benwing2 (talk) 14:55, 19 May 2018 (UTC)
I have removed the block. I find the vote to be an absurd waste of everyone's time, Wyang's deletion of the vote without discussion to be inappropriate, and his block to be baseless. —Μετάknowledgediscuss/deeds 17:09, 19 May 2018 (UTC)

Proficiency as a prerequisite for contributionEdit

Voting and the strength of the arguments and evidenceEdit

Some people say that in a vote, the strength of the arguments and evidence all too often do not prevail, and that, instead of votes, there should be discussions only, to be closed based on the strength of the arguments presented. Contrary to this, I submit the following thesis:

The strength of argument and evidence is more likely to prevail in a vote combined with discussion than in a discussion that is closed by a single closer based on their assessment of the arguments and evidence.

I do not have a formal empirical proof of the above, but the following consideration is suggestive. In a vote, all the voters have the opportunity to consider the arguments presented, and vote accordingly. In order for a 2/3-supermajority vote to pass a bad decision, 2/3 of the voters have to fail in their assessment of the arguments and evidence. By contrast, in order for a single-closer discussion to pass a bad decision, it suffices that the single closer fails in their assessment. Thus, votes are much more robust. An objection can be made that many people are not careful enough, and possibly do not have the talent. Sure, and many of such people are all too eager to close a discussion based on their assessment of the arguments; thus, the existence of people's weaknesses does not make single-closer discussions any more reliable. The quote attributed to Bertrand Russell applies: "The trouble with the world is that the stupid are cocksure and the intelligent are full of doubt."

Another point is that of matter of taste. In decisions that are a matter of taste, there does not necessarily exist such a thing as a strong argument. With a single-closer discussion, the single closer's taste prevails, for no good reason that I can discern; in a vote, the tastes of multiple people are considered.

Sure enough, as for the strength of argument and evidence, not all votes are equal. Votes in which comments and questions to posters are not allowed are weaker than votes in which discussion and inquiries are allowed. Similarly, votes where people are encouraged to state their rationale are stronger. Furthermore, votes where enough time is ensured for discussion before the start of the vote are stronger.

--Dan Polansky (talk) 11:05, 27 May 2018 (UTC)

lol you're making the classic mistake of thinking that other human beings are people Equinox 11:11, 27 May 2018 (UTC)
Ok, I'll use the phrase "sentient beings" next time around, to include e.g. centaurs :). --Dan Polansky (talk) 11:23, 27 May 2018 (UTC)

Another objection: The discussion closers could be members of a selected group of people, such as admins. These could be chosen to have increased ability in assessing the strength of argument and evidence. One set of problems with that is the reliability of the process of selection, who would be allowed to participate in such a selection, that admins are chosen also by other criteria than the ability to assess arguments and evidence, and how unwiki that really would be. But even if we put these kinds of problems aside, the fundamental distinction between a single closer and a multitude remains: a vote in which only the qualified elite or citizens in the sense of Ancient Greece are allowed to vote whereas the mere non-citizens are only allowed to present arguments and evidence, is still stronger and more robust than a single-closer discussion. --Dan Polansky (talk) 11:23, 27 May 2018 (UTC)

Real response: 1. Voting systems have been studied by thousands of people for centuries and we are probably not going to invent any new kind of system that works better. So let's forget the specific mechanisms of voting. 2. The same applies to your comment about the elite versus the plebs (admins versus IPs or whatever): it's probably true but we are best served by looking at existing literature on group decision-making. 2b. This is really about class systems, and whether anyone's vote can have MORE WEIGHT than someone else's. I won't dip my toe in that shitstorm. 3. My CRAZY PERSONAL FEELING is that the main biasing factor is actually "interference" (I don't know the correct psychological term): that is to say, when we do have a vote, you don't just go in blind and vote for what you think. You look at the entire page and the discussion, and maybe that will nudge you a bit. I listen to an Internet radio a lot where you can rate tracks 1 to 5 stars and I very often think "ugh this is a 2" and then I see the average vote is a 4 and I think "ahh well maybe it's a 3". Peer pressure! (Also in the demoscene there is a phenomenon known as "namevoting" where people will vote bad shit very high if it's made by a famous nerd.) Go figure. Equinox 14:08, 27 May 2018 (UTC)
Thank you. Let me note that my point was not to delve into merits of demotic/plebsist vs. elitist voting; rather, my point was that the issue of vote vs. single closer is orthogonal to that, and therefore, the whole demotic vs. elitist divide can be factored out of the discussion. On another note, a literature review would be the proper academic exercise. Nonetheless, there is a beauty in presenting a simple, yet persuasive (to some) argument in contrast to literature review. Of course, one factor is that I can conjure up an argument real fast, at the risk of being wrong, whereas literature review is a lot of work :). Furthermore, literature review is no guarantee in drawing the right conclusions. Some people, reviewing the Arrow's Theorem (e.g. in SEP), draw the wrong conclusions from it about the merits of voting. Yes, each method of aggregating preferences leads to anomalies and problems, yes, each axiomatization of arithmetic leads to problems, yes, each system of vision is vulnerable to certain visual fallacies. What of it, abandon all systems with weaknesses? The various negative results show that, for some tasks, there are none such systems in the design space. --Dan Polansky (talk) 18:51, 27 May 2018 (UTC)


This user likes to add images as a decoration to add "a certain style". I'm bringing this to your attention because you care about this more than I do, and the last time I reverted them, it felt like shooting a puppy in front of a 4-year-old... Chuck Entz (talk) 15:38, 27 May 2018 (UTC)

Thank you. I removed two images in diff and diff. A related discussion is at #Removing off-topic images, and a related vote is Wiktionary:Votes/pl-2018-04/Image policy‎, still running. More images added by the user could be removed, but I'll wait for later; perhaps someone will help. --Dan Polansky (talk) 15:56, 27 May 2018 (UTC)

Translation hub label or indicationEdit

Discussed here: Wiktionary:Beer parlour/2018/May#Translation hub label. So far, no support for using {{lb|en|translation hub}}. I pointed out we could use indication similar to hot word. I strongly dislike putting "(This entry is here for translation purposes only.)" or "(This entry is a translation hub.)" on the definition line via {{translation only}}; it is not a definition, not even a non-gloss definition. It is the real abuse, not the label; it is labeling without a label, and a wordy one too, with unnecessary "This entry is a". --Dan Polansky (talk) 06:40, 2 June 2018 (UTC)

Vote pass thresholdEdit

The threshold for a vote passing is not specified anywhere, and is a long-term open issue for the English Wiktionary. It is usually not much of a problem since votes that have at least 70% of supports to support+oppose and have decent participation are usually closed as passed without any ado.

Various thresholds have been mentioned in discussions in Wiktionary forums, including 80%, 75%, 70%, 2/3 and even 60%. Keyword: two thirds.

I support that, for many kinds of votes, 2/3 of supporters from support+oppose count is the threshold to be used, or even somewhat lower, provided there is significant participation, and provided the vote runs reasonably long, which for a policy vote is at least a month. Thus, 10:5:1 (support:oppose:abstain) would be a pass. With low participation (e.g. 4:2:1), I would be uncomfortable closing 2/3 vote as a pass, while 5:1:1 would be a pass, I feel. Some votes may be more of a constitution-changing character, as it were, and for these, 2/3 could be too low.

Votes closed at a threshold equal to or lower than 2/3:

There may be other ones.

Related Wikipedia articles include W:Supermajority and W:Constitutional amendment.

Another way of measuring the threshold is as the ratio of support to oppose. Thus, 2/3 of supports is equivalent to 2:1 ratio of support to oppose, 60% of supports is equivalent to 1.5:1 ratio of support to oppose, and 50% of support is equivalent to 1:1 ratio. The ratio gives a different picture of differences between various thresholds: while 2/3 is 11% higher than 60%, 2:1 is 33% higher than 1.5:1. --Dan Polansky (talk) 10:06, 2 June 2018 (UTC)

X language entriesEdit

I often support keeping "X language" entries even when they can be construed as sum of parts. Some past discussions:

A current 2018 discussion is at WT:RFDN#Names of languages in Belarusian, e.g. on беларуская мова.

My keeping principle: keep if the "X language" forms are more common than "X" alone in reference to the language. Avoid deleting the most common way of referring to a language. The sum-of-parts rationale has to exist in some form, but it should also be overriden on a case-by-case basis to make the dictionary user-friendly.

Also consider that when X is an adjective, the use of X to refer to the language is just an abbreviation of "X language". This case is different from Czech čeština, a noun that has "český jazyk" as a synonym.

--Dan Polansky (talk) 09:04, 10 June 2018 (UTC)

Hyphenated terms and sum of partsEdit

Hyphenated terms are repeatedly nominated for deletion in RFD as sum of parts. Such terms include disease-free and HTML-based, currently in WT:RFDE, and later at Talk:disease-free and Talk:HTML-based.

Some argue that hyphenated terms are single words and should be thereby automatically kept if attested.

A hyphenated term may be protected by WT:COALMINE; thus, man-made is protected via existence of manmade, which is less common per man-made, manmade at Google Ngram Viewer. Another protector may be WT:LEMMING.

Past deletions include Talk:Godzilla-sized, Talk:dark-colored, Talk:download-only, Talk:Japan-only, Talk:testosterone-fueled, Talk:71-year-old, Talk:twelve-year-old, Talk:Finnish-Canadian, and Talk:US-based.

Past keeps include Talk:ever-watchful, Talk:ex-pilot, Talk:faster-than-light, Talk:Swedish-Canadian, Talk:H-shaped, Talk:high-quality, Talk:high-resolution, Talk:multi-word (coalmine), Talk:non-Japanese, Talk:oft-used (coalmine), Talk:out-Herod, Talk:post-Soviet, Talk:re-explain, Talk:rod-shaped (coalmine), Talk:seventy-first, Talk:short-legged (coalmine), Talk:sword-bearing (kind of coalmine), Talk:two-wheeled, and Talk:well-rested (coalmine).

Terms never nominated include many of those in Category:English words suffixed with -free, like alcohol-free and caffeine-free.

--Dan Polansky (talk) 10:17, 24 June 2018 (UTC)

A special case is that of a hyphenated form that is the attributive form of a space-separated non-SOP term, e.g. transitive-verb corresponding to transitive verb. Deletion discussions for attributive nouns include Talk:transitive-verb, and Talk:cookie-cutter; and also Talk:transitive-verb#All English attributive forms (with hyphens) of noun phrases as well. Yet another discussion is now at WT:RFDE#geographical-area, later at Talk:geographical-area. --Dan Polansky (talk) 13:23, 2 July 2018 (UTC)

Title for collapsible boxes for Related terms and the likeEdit

Czech inflection table coverageEdit

The number of inflection tables entered can be estimated from the following searches:

  • Verbs: insource:/\{\{cs-conj/: 1690 hits
  • Nouns: insource:/\{\{cs-decl-noun/: 4307 hits
  • Adjs: insource:/\{\{cs-decl-adj/: 1747 hits
  • Sum total: 7744

The coverage of inflection tables expressed as a percentage is as follows:

- Verbs: 53% = 1690 / 3192 Category:Czech verbs
- Adjs: 27% = 1747 / 5883 Category:Czech adjectives
- Nouns: 25% = 4307 / 17478 Category:Czech nouns
- All 25% = 7744 / 30395 Category:Czech lemmas

--Dan Polansky (talk) 12:11, 28 September 2018 (UTC)

Not bad, DP. I remember helping out with some of these many years ago, when I pretended I was knowledgeable in Czech in order to gain adminship. Lol, I even ran a Czech bot. Anyway, I remember {{cs-decl-noun-auto}} being useful back then. How good is that template these days, anyway? --XY3999 (talk) 12:21, 28 September 2018 (UTC)
It actually is pretty bad, I'd say. The thing is, I consider inflection to be of secondary importance and rarely add inflection tables. As a result, the credit for the inflection table coverage mostly goes to other people, including some anons. The credit for the Czech verb coverage goes to Jan.Kamenicek to a significant degree; he is responsible for {{cs-conj-forms}}, which I find to be a very decent template, and he made significant rollout of the template in the mainspace. The auto decl template for nouns is pretty good but not perfect. --Dan Polansky (talk) 12:35, 28 September 2018 (UTC)
I still would like to see someone run a Czech bot. Romanbot (talkcontribs) and Romanb (talkcontribs) were the usernames used for the Czechbot by that rogue admin guy (Wonderfool, or something). Let's bring em back --XY3999 (talk) 14:31, 28 September 2018 (UTC)
What is this magical search term metalanguage that I've never heard of, and where can I learn more about it? Allahverdi Verdizade (talk) 11:20, 13 December 2018 (UTC)
A documentation about insource is in Help:CirrusSearch#Insource, The language used between /.../ are B:Regular expressions; you can also use 'insource: "word1 word2"', without regular expressions. --Dan Polansky (talk) 07:56, 15 December 2018 (UTC)

Inclusion of senses for buttons and keysEdit

Discussions about deletion of senses for buttons and keys:

--Dan Polansky (talk) 09:27, 7 October 2018 (UTC)

Common misspellings vs. common typosEdit

Some argue that we should exclude all attested typos, common or not, while possibly keeping common misspellings that are not typos. However, the rationale for keeping common non-typo misspellings applies to common typos as well: the chance that someone will run into them in text is relatively high. WT:CFI#General rule starts with "A term should be included if it's likely that someone would run across it and want to know what it means." That's the idea.

For completeness, WT:CFI#Spellings is currently specified in terms of misspellings, not typos: "Rare misspellings should be excluded while common misspellings should be included".

In User Talk:Dan Polansky/2013#What is a misspelling, I proposed to use a frequency ratio criterion to determine whether to rank something as a misspelling. That criterion is easy to administer with the help for Google Ngram Viewer, for multiple languages. I really do not see what possible advantage there could be for the user in our introducing a metaphysical distinction harder to detect in addition to the quantitative one. --Dan Polansky (talk) 09:38, 2 December 2018 (UTC)

  • I disagree. I think common misspellings are different because some people, even without seeing the word in print, honestly but mistakenly believe that the misspelling is the correct spelling. This has to be the case, otherwise these misspellings wouldn't make it into print in the first place. bd2412 T 01:54, 6 December 2018 (UTC)
    • The frequency ratio criterion only deals with things found in print. Thus, if we believe that "otherwise these misspellings wouldn't make it into print in the first place", the forms having a high relative frequency in print would have to be non-typo misspellings anyway. The notion of frequency is simple, easy to detect and does all that we need, as far as I can see. --Dan Polansky (talk) 13:15, 8 December 2018 (UTC)
Return to the user page of "Dan Polansky".