Category talk:Latin declension templates

Organizing the templates
As a new contributor, I found the templates within this category to be somewhat scattered and disorganized. Therefore, I went ahead and organized the templates. The organization scheme should be pretty obvious:


 * '1' through '5' group the templates for substantives of a given declension.
 * 'A' groups the adjective templates. It might make sense to put 3rd decl. adjectives under 'B'(?)
 * 'P' groups the pronouns and related adjectives.
 * '~' groups the deprecated templates in this category.

I hope this helps picking out the correct template for a given word. maikxlx (talk) 16:57, 20 December 2008 (UTC)

Deprecated templates
I am placing unmistakably deprecated templates within this category under the header "~" and putting the following tag in the  area of the template.

''' This template is deprecated. ''' See Category_talk:Latin_declension_templates for ore discussion.

The following two templates are obsolete, unused except for one link to one person's user talk page, and therefore can be safely deleted:
 * Template:la-decl-3rd-N-IUM
 * Template:la-decl-3rd-INEUTI

Proposal to rework the templates
Hello! In the last few days, I have been discussing the Latin declensional template situation with EncycloPetey. The current templates seem inadequate in several ways:


 * 1) Every Latin noun table links to its own page.
 * 2) It's impossible to add a footnote or modify a cell in any of the currently used paradigmatic templates in order to indicate a variant form or defect in the paradigm.  Since Latin nouns are rife with variants and defects, this means a large number of nouns can't be declined using the appropriate paradigmatic template; alternately, the Wiktionary entry is simply inaccurate.
 * 3) The singular-only and plural-only entries require their own templates.  This theoretically triples the number of templates.
 * 4) Showing the locative case involves another duplication of templates.  This theoretically doubles the number of templates.

I have designed a new type of Latin declensional template which fixes (1) and addresses (2)-(4) through optional named parameters. Please see how it works here and let me know what you think. These templates are backward compatible with current template calls, except possibly for the 3rd declension, which I would like to discuss as a separate issue.

Here is my proposed plan of attack:

Refit most decl. 1, 2, 4, 5 templates
This simply means upgrading the current templates. No template calls are affected by this change, but all of the advantages of the new template design become available.

Deprecate all loc, sg-only, pl-only templates
Deprecate the following templates and rework the template calls ASAP (# = number of current uses):
 * Template:la-decl-1st-loc 23
 * Template:la-decl-1st-p-loc 2
 * Template:la-decl-2nd-loc 2
 * Template:la-decl-2nd-N-loc 7
 * Template:la-decl-3rd-loc 2

Rework the decl. 3 templates
Create new the following templates:
 * Template:la-decl-3c-MF
 * Template:la-decl-3c-N
 * Template:la-decl-3i-MF
 * Template:la-decl-3i-N

Deprecate the following templates and rework the template calls ASAP:
 * Template:la-decl-3rd-INEUT 5
 * Template:la-decl-3rd-PAR-navis 3
 * Template:la-decl-3rd-PAR-ignis 8

Deprecate the following templates and rework the template calls as time permits:
 * Template:la-decl-3rd-PAR 161
 * Template:la-decl-3rd 340
 * Template:la-decl-3rd-N 46

Rationale for creating new decl 3 templates
1. The parameter order of the current decl 3 templates is flawed. With a proper parameter order, there would be never a reason to provide more than two parameters in the case that the nom. sg. & stem contains no long vowels. See examples of the decl. 3 template I created here for an improved parameter order.

2. Some current decl 3 template names are flawed or inconsistent. Most entries under "la-decl-3rd-PAR" are not parasyllabic. "la-decl-3rd-INEUT" does not follow the convention of suffixing "-N" to templates for neuters. This adds confusion to the system. The names I am proposing seem much clearer.

I'd like to get these things right now, while we are reviewing these things in detail. Bear in mind that deprecating a template does not mean deleting it or breaking it - on the contrary, we can upgrade them immediately to the new standard and continue to maintain them without affecting anything else. All deprecated templates except for two or three can be obsoleted very quickly.

If this proposal meets general approval, I'd like to get started with refitting ASAP while I have the time. I will leave aside the decl 3 templates until everyone has a chance to review the proposal.

Discussion
Some of your proposal is lacking in explanation of what is intended. You propose deprecating the locative templates, but have not indicated what would replace them. The proposed renaming of the 3rd declension templates is at odds with some of your proposal. All of the other declension tables use -1st, -2nd, -4th, etc., but you are proposing that we change the names of the 3rd declension templates to use -3c / -3i instead of -3rd. I don't follow this logic. The current system is more consistent about the names. And finally, as I said before, I disagree with the reordering of the 3rd declension parameters. I disagree in part because the current system is consistent with the order used by other Latin templates; it would be a poor choice to have one group of templates using a different ordering from all the others. There are ways to make the current order work within your new templates. --EncycloPetey 05:22, 22 December 2008 (UTC)
 * (maikxlx:) To address EncycloPetey's response first:


 * Some of your proposal is lacking in explanation of what is intended. &mdash; I know that my explanation is lacking. I sketched an rough outline of my proposal; the rest will have to be covered in discussion.--maikxlx


 * You propose deprecating the locative templates, but have not indicated what would replace them. &mdash; I propose replacing the locative-displaying templates precisely with the regular template of the appropriate paradigm, and then invoking named parameters to trigger a change in the display. I also propose that any sg-only or pl-only templates (currently needed, but not yet written) are handled the same way.  Given the potential occurrence of variations +/-loc, +/-sg-only, etc. in combination in several paradigms, I would argue that it's more efficient to handle variations through parameters, rather than writing a full-blown template for every special case. --maikxlx
 * OK, that makes sense. Could you create an example for one locative to demonstrate how this will work?  If you could do the same with a regular masculine 2nd declension, it would allow more people to see what you're planning.  I can read the template syntax and the Latin, but most other users here can't read both without intense effort.  We have some coding-experts here that might be able to streamline the new template code. --EncycloPetey 21:42, 22 December 2008 (UTC)
 * I did add some examples here yesterday that you might have missed. I will say more about the locative and number options later this evening, though, that should make my ideas on this issue clear.  I will set up a few 2nd decl. examples for you too. --maikxlx (talk) 22:58, 22 December 2008 (UTC)
 * I saw the eamples, but wanted you to include some here for those who do not already know exactly how the templates (and Latin) work. --EncycloPetey 19:53, 24 December 2008 (UTC)


 * In other words, I propose an organizational principle for Latin templates of exactly one template per paradigm (and exactly one paradigm per template). I propose treating any variation of a given paradigm - whether it be alternating forms, defects of declension, locative case display, or suppressing the sg or pl - as special cases of the same paradigmatic template handled by invoking the appropriate optional named parameters.  The beauty of this system is that we can add parameters if needed as we go along, without affecting anything in production. --maikxlx
 * I think I follow, but I assume you plan for a single master 1-column and single master 2-column that all the templates call to, yes? That way, the style can be set in just two places, making it easier to update in the future. --EncycloPetey 21:42, 22 December 2008 (UTC)
 * Actually, the way I designed it is like this: There is a single master template for all nouns.  There are three supporting "style" templates that contain the style information:  One for the ordinary cells, one for the row headers, and one for the column headers.  Under this design proposal, you could literally change the style of every Latin noun entry at en.Wiktionary with a few key strokes.  --maikxlx (talk) 22:58, 22 December 2008 (UTC)


 * The proposed renaming of the 3rd declension templates is at odds with some of your proposal. All of the other declension tables use -1st, -2nd, -4th, etc., but you are proposing that we change the names of the 3rd declension templates to use -3c / -3i instead of -3rd. I don't follow this logic. The current system is more consistent about the names. This is a minor issue.  If we were writing these templates from scratch, I'd propose the naming convention -1-, -2-, -3-, etc., as the strings -1st-, -2nd- and -3rd- add little but extra letters.  However for consistency with the existing system, maybe we can also name them this --maikxlx:
 * Template:la-decl-3rd-c-MF
 * Template:la-decl-3rd-c-N
 * Template:la-decl-3rd-i-MF
 * Template:la-decl-3rd-i-N


 * Or this even:
 * Template:la-decl-3rd-cstem-MF
 * Template:la-decl-3rd-cstem-N
 * Template:la-decl-3rd-istem-MF
 * Template:la-decl-3rd-istem-N


 * Something like the first listing looks better to me, but I don't think the uses of "c" or "MF" are necessary. The 2nd declension templates explicitly call "N" for neuter entries, but the default is for masculine.  We could do the same in the 3rd declension and have M/F and "c" be the default, without specifying it in the name.  I have no argument with using "i" instead of the current "PAR", since I think more authors use "i-stem" than "parasyllabic" anyway.  I therefore expect more contributors would be able to quickly learn the system with your proposed use of "i" in the template names. --EncycloPetey 21:42, 22 December 2008 (UTC)
 * Well, I could go along with what you say here. That would coincide with existing templates "la-decl-3rd" and "la-decl-3rd-N".  If I were allowed only one change, it would be to get rid of "PAR".  It's at once cryptic and inaccurate. For consistency let's change "INEUT" to "-i-N" too.  The rest of the names are not an issue.  --maikxlx (talk) 22:58, 22 December 2008 (UTC)


 * And finally, as I said before, I disagree with the reordering of the 3rd declension parameters. I disagree in part because the current system is consistent with the order used by other Latin templates; it would be a poor choice to have one group of templates using a different ordering from all the others. There are ways to make the current order work within your new templates. &mdash; The current 3d order is not consistent with the 1/2/4/5d order; the 3d order is 1=nomsg-wo-macron, 2=nomsg-w-macron, 3=stem-wo-macron, 4=stem-w-macron; the other decl. orders are all 1=stem-wo-macron, 2=stem-w-macron. If the 3d order were consistent with the others, it'd be stems first, and _then_ nomsg.  I can certainly accommodate the current 3d order, but I want to take this one-time chance to argue in favor of the efficiency of specifying the nom.sg. only once in the many cases when no macrons occur.  It's annoying to be forced to write things like   in cases when   ought to be entirely sufficient. All I am asking is that you consider it.  --maikxlx (talk) 11:18, 22 December 2008 (UTC)
 * Yes, the current system is consistent with the other templates. Please note that I do not simply mean the other declension table templates; I am referring to all the various Latin language templates, including those for the inflection line and for other parts of speech.  The usual order for the other templates is to code the stems in pairs with their macron-bearing forms next to the forms without macrons.


 * Additionally, while there is some efficiency in re-ordering the code, there is also more opportunity for input error. The current parameter order is more likely to reveal mis-coding by ommision then the proposed order would.  This is something I have seen more than once in other languages' templates as well.  If putting in two parameters makes the 3rd-declension templates "work" (i.e. something appears in the tables), then an editor is less likely to realize a mistake was made in parameter input.  The 3rd declension templates are the only ones that require two different stems—the others require only one—and even seasoned editors sometimes forget this fact.  The current template order forces an editor to remember because a missing 3rd parameter causes an error in the tables, which is a visual reminder that these templates are different. --EncycloPetey 21:42, 22 December 2008 (UTC)
 * These comments make your position clearer. I will comment in the dedicated subsection. --maikxlx (talk) 22:58, 22 December 2008 (UTC)


 * I am a bit confused as well. You propose many changes, at least a few of which I agree with, but it would be helpful if we could break this down a bit.  First of all, can you give us a concrete example of the reordering bit?  I'm not following how reordering any parameters would make a difference.  I strongly agree with adding the ability to specify forms manually.  If Latin's at all similar to Ancient Greek in this, as I suspect it is, there will be many uses for such specification abilities.  However, I think it a bad idea to append notes inside the cells.  The problem is simply one of space.  I suggest using an approach similar to what I'm using with grc templates, where there is a notes section under the inflection table.  See  (and its inflection template Template:grc-decl-3rd-N-prx-con-εος) for an example of how this works.  The beauty of this is that extensive commentary can take place if need be (which is admittedly rare...for now).  Also, I suggest putting the templates in a hidden box.  I know EP has reservations about them, but I wonder how many people really have difficulty in figuring out how to open them.  This saves a lot of space in an entry, which becomes especially important when multiple tables come into play, as in {term|ἔτος|lang=grc|sc=polytonic}}.  Finally, I agree with EP that 3rd should remain in the name of the third declension templates, but there is definite merit to your renaming scheme.  Could we do la-decl-3rd-C-MF or something like that?  That's all I can think of for now.  I'll continue to investigate your new templates more fully, and see if I can figure out some of the things which confuse me.  -Atelaes λάλει ἐμοί 07:14, 22 December 2008 (UTC)
 * I am indifferent to the hidden box idea, but I am willing to add them to the design if they are desired.--maikxlx (talk) 13:19, 22 December 2008 (UTC)
 * As Atelaes has noted, I dislike collapsible boxes for nouns and adjectives. Verbs need them because the tables are huge, and Ancient Greek and Hungarian need them for other parts of speech because there can be multiple tables.  For Latin, I'm not convinced at this point that nouns and adjectives would benefit.  I do think the table templates ought to have a Header row identifying the declension pattern and linking to the appropriate Appendix(-ces) or appendix section.  In the past, this was done by placing an additional template ahead of the table, but it makes more sense to combine the two calls into a single template.  I already did this for the verbs, and think it only makes sense to do so for the nouns (and other parts of speech with tables) as well. --EncycloPetey 21:47, 22 December 2008 (UTC)
 * As far as the other pattern info, we already have the "la-noun" template that gives the declension and a link to the Appendix. However if truly desired, we can also add info to the decl templates easily enough, and delete the intro messages.  --maikxlx (talk) 22:58, 22 December 2008 (UTC)
 * Yes, the inflection line links to the Appendix, but not to the relevant section of the appendix (and it shouldn't). The Inflection section is designed to carry all relevant information about the inflection, which the inflection line is a short summary of key points.  Since the Inflection section should have all the relevant information, it seems best to identify the particular inflectional pattern there (1st, 1st-Greek, 3rd-i-stem, irregular, 2nd and 4th, etc.) for each noun. --EncycloPetey 19:56, 24 December 2008 (UTC)


 * Ok, I think I've got it all figured out now. You've certainly put together an impressive array of new features, and you have my congratulations for a job well done.  A response to EP, to answer his questions, if I may.  What is being proposed for the locative templates is this: Currently the new-style templates access another template, which contains all the formatting info (what colors the table is, what the rows are titled, etc.).  What is being proposed is to have a few options for this secondary template.  So, instead of accessing la-decl-1st-loc, as is currently done, you access la-decl-1st, and tell it to access Latin-decl-loc (I think they should actually be named la-decl-blank-full and la-decl-blank-loc respectively, but that's about as minor an issue as you could hope for).  Thus, la-decl-1st always knows what the locative forms are, but it usually doesn't display them, because Latin-decl doesn't have spots for them.  Same with plurals only and singulars only.  You're accessing the same template, but telling it to feed its information through a different blank template.  This is how we do it with the grc templates, and it's really the way to go.  I've already said my bit about naming the third declension templates, and I stand by it still.  As for the ordering of the third declension templates, I tend to side with EP.  I agree with EP that you'd really want to reorder all the templates, for consistency if you're going to do this, which is a whole lot of work for, as I see it, no gain.  I'm presuming that we generally want macrons, and so the current order makes more sense, as it encourages inserting them.  However, you can still use the templates if you don't know the macroned form, you simply leave parameters 2 and 4 blank.  It should still work properly.  -Atelaes λάλει ἐμοί 09:02, 22 December 2008 (UTC)


 * Also, Maikxlx, I've placed a note at User talk:Maikxlx/Template:la-form. I didn't know if you had that watchlisted or not.  -Atelaes λάλει ἐμοί 09:05, 22 December 2008 (UTC)


 * Atelaes, could you please direct me to your previous comments about naming the third declension templates? I am curious as to your opinion. --maikxlx (talk) 12:21, 22 December 2008 (UTC)


 * It's in the first paragraph, and reads Finally, I agree with EP that 3rd should remain in the name of the third declension templates, but there is definite merit to your renaming scheme. Could we do la-decl-3rd-C-MF or something like that?.  Basically, exactly what you've already proposed in response to EP. :-)  -Atelaes λάλει ἐμοί 12:31, 22 December 2008 (UTC)

Decl 3 reordering Issue
I want to respond to one of Atelaes's comments but at this point I think that it's helpful to hash out each issue in a separate section. Atelaes asked for an explication of my 3d reordering proposal. My basic objection with the current 3d ordering is that it places optional parameters in front of required patterns. The following is (IMO) the natural and logical approach (maikxlx (talk) ):
 * 1) Paradigms in which the nom-sg. is predictable from stem: 1=stem-wo-macron (required); 2=stem-w-macron (often unneeded)
 * 2) Paradigms in which the nom-sg. is unpredictable: 1=nomsg-wo-macron (required); 2=stem-wo-macron (required); 3=nomsg-w-macron (often unneeded); 4=stem-w-macron (often unneeded)

What we have now is (maikxlx (talk) ):
 * 1) Paradigms in which the nom-sg. is predictable from stem: (same as above)
 * 2) Paradigms in which the nom-sg. is unpredictable: 1=nomsg-wo-macron (required); 2=nomsg-w-macron (often unneeded); 2=stem-wo-macron (required); 4=stem-w-macron (often unneeded)

In the more complicated paradigm type in the current system, you have to specify the nomsg-w-macron parameter even if the thing doesn't have macrons. No matter how you slice it or dice it, you're going to have two types of paradigms. The difference is that the current system forces the editor to sometimes specify unneeded parameters, and the other does not. --maikxlx (talk) 13:32, 22 December 2008 (UTC)


 * I have responded to this above, but because the discussion is likely to be wordy, I have duplicated my comments here:


 * The current system is consistent with the other templates; your proposal is not. Consistency between the various templates is a great boon to editors, but inconsistency is a terrible bane.  (Users won't notice the difference, so only editors need be considered here.)  Please note that I do not simply mean the other declension table templates; I am referring to all the various Latin language templates, including those used for the inflection line and for the other parts of speech.  The usual order for the other templates is to code the stems in pairs with their macron-bearing forms next to the forms without macrons.
 * I was almost going to say you might have a point here, until I actually started looking around and saw this:

[...]


 * Surely consistency would demand that the verb citation template and the verb conjugation template have the same parameter structure, especially when one takes six parameters and the other takes eight! Not to mention that most of those forms are totally predictable and could have been built by the template!  Who is the poor suffering editor who's adding these entries anyway?  Whoever he is, I admire his fortitude!  Now, I agree that consistency is important, but consistency with what?  Instead of talking about consistency with the current kludges, why not talk about consistency with a better planned system?--maikxlx
 * The verb conjugation tables are the one major exception. I dislike the difference and become confused about the syntax occasionally myself, even after working with them all this time.  While the form are predictable for completely regular 1st conjugation verbs, most verbs are not that regular, and there are 1st conjugation mostly-regular verbs for which those endings are not used.  The current template supports a greater number of 1st conjugation verbs than it would if the most-regular endings were forced to be used. I disagree with you about the current system having kludges.  The endings can be predicted as you say, but not for all verbs.  That's why Latin dictionaries list the four principal parts.  The 1st conjugation is more often regular than the other conjugations, but not entirely.  The choice for the 1st conjugation template then is: (a) have the template build the forms (which would not satisfy all verbs and would have it function differently from all the other verb conjugation tables), or (b) use what we have, which is consistent with the other tables and therefore easier to learn and remember. --EncycloPetey 20:08, 24 December 2008 (UTC)


 * As an aside, let me say something positive: The 1st conj verb template has a good parameter structure--it wisely separates the macroned and macronless stems instead of intermingling them.  Also, I think that putting the macroned stems in front of macronless stems makes more sense, arguably more logical than the way I might have had it.  Well done; this should be the parameter order for the whole system.  Now, if you wanted to get maximally efficient, you could easily have the template build the perfect stem and the supine stem from the present stem whenever the latter are not specified, as the most common pattern is entirely predictable.  In other words, a lot of 1st conj verb entries could be built from two parameters instead of six (or eight in the case of the citation template).  But, I digress.--maikxlx
 * Having the template build the forms would add another layer of unnecessary complexity into the template. The current system is designed to reduce error rate by requiring the stems to be added explicitly.  There is an addition reason for this method in the no-passive versions of the templates, where there may not be a supine or participle form for a particular verb, and so absence of that form indicates it is lacking those forms, so those cells are empty.  I would rather that we didn't use different approacxhes for different templates in the same template group. --EncycloPetey 20:08, 24 December 2008 (UTC)


 * Additionally, while there is some efficiency in re-ordering the code, there is also more opportunity for input error. The current parameter order is more likely to reveal mis-coding by ommision then the proposed order would.  This is something I have seen more than once in other languages' templates as well.  If putting in two parameters makes the 3rd-declension templates "work" (i.e. something appears in the tables), then an editor is less likely to realize a mistake was made in parameter input.  The 3rd declension templates are the only ones that require two different stems—the others require only one—and even seasoned editors sometimes forget this fact.  The current template order forces an editor to remember because a missing 3rd parameter causes an error in the tables, which is a visual reminder that these templates are different. --EncycloPetey 21:42, 22 December 2008 (UTC)


 * I don't believe that typing extra, often redundant input is going to reduce many errors. Anyway, as we might not reach agreement on this tonight, I'm going to work on the locative-variant aspect of my proposal.--maikxlx (talk) 01:30, 23 December 2008 (UTC)

Locative and Number Variants - 2nd declension examples.
Please review my latest examples here and discuss my proposal in this section on this page.--maikxlx (talk)

The 2nd declension m (-us) paradigm was a great choice for examples because it requires no fewer than SIX number and locative variants of the basic paradigm -- notice the apt selection of nouns. We do not have all of these templates currently, nor should we create them; the flexibility of my template design takes care of this.--maikxlx (talk)

Please note that I am offering a choice between two proposals - we can use either one optional named parameter to specify variants, or two. Please read this for details. --maikxlx (talk) 14:06, 23 December 2008 (UTC)


 * I have a slight preference for the single parameter scheme. Ultimately, having to add one more parameter is not a big deal, but it is a little bit more work.  Admittedly, the single parameter scheme is more prone to result in confusing amalgamations.  -Atelaes λάλει ἐμοί 16:40, 23 December 2008 (UTC)

Where
Templates sure have become confusing since I was last an active Wikipedian in the 2000s.

Where on earth are the actual data for an irregular word like deus? The various forms (deum, deōrum, dīvom…) are neither auto-generated (because they can't be) nor manually entered into the article in question. They are squirrelled away in some non-obvious location for reasons. Correctrix (talk) 13:35, 10 October 2021 (UTC)