hu-suffix

The changes you made in {{hu-suffix}} created an unwanted side-effect. In some cases, we need to use multiple suffixes in a single etymology, but we want to categorize the entry only by the last suffix. With the new change, even the intermediate suffixes produce a category which is not correct. See rugalmas. Can you add this functionality to {{suffix}}? If not, please restore the original hu-suffix. I don't understand why it was necessary to replace the Hungarian-specific template. Not to mention that this was done without discussing it with the Hungarian editors. Is there a template policy that says FL entries must all use the general templates?

There's no policy, but it seemed to make more sense to have it written in terms of the generic template, if at all possible. I wasn't aware of that specific case though so I guess it will need to be undone.

But I do wonder how it works with rugalmas. If you just attach the first suffix, is that not a valid word?

The word rugalom is valid archaic noun, rugalmas is an adjective. The Hungarian suffix categories contain the PoS in their name because some of the suffixes have multiple purposes, so we want to separate the derived terms by PoS. The {{suffix}} template allows only one PoS parameter and would put the headword that has many suffixes into different categories marked with the same PoS. Another consideration is that sometimes the ety has to indicate linking vowels which are not suffixes, so categorization would make no sense. See nézeget as an example. Hungarian words can have a number of suffixes appended after each other. E.g.: szabálytalanság = szab + -ály + -talan + -ság. If I use {{suffix}}, the noun headword would appear not only in Hungarian nouns suffixed with -ság, but also in the categories Hungarian nouns suffixed with -ály and Hungarian adjectives suffixed with -talan. Too much clutter, the result would not be clean.

But we don't normally show the complete morphological/etymological breakdown of words. Instead we only show the "outermost" derivation, the one that was applied last. So for rugalmas it would just be rugalom +‎ -as and for szabálytalanság szabálytalan +‎ -ság. Compare this for example to unevenness, which is not un- +‎ even +‎ -ness but just uneven +‎ -ness.

In the case of szabálytalanság, you're right. I would not divide it up. I used this word only to illustrate the number of suffixes. In the case of rugalmas, I wanted to show only the modern stem and not the archaic word.

But is that the real etymology? I suspect that when the word was first created, it was from rugalom when it was still in normal use.

This is a complex issue in Hungarian. First the facts: rugalom seems to have existed but I have no information whether

• rugalmas is derived from rugalom or
• rugalom was a back-formation from rugalmas

(and whether such a distinction conceptually really makes sense, but more on that further down)

The references that we (with Panda10) have current access to don't mention rugalom in the etymology of rugalmas, just rúg. According to the (probably very much outdated) 1902 Dictionary of the Hungarian language reform:

Rugalmas, Helmeczy (Törv.). Megvan már 1829-ben (Pák., Vad. 2:193). A Közhasznú Esm. tárában Nyiry István az elasticitást kelékenységnek akarta nevezni. Hogy Barczafalvi (Szigv.) és Bugát (Tsch.) honnan vehették az ermetzes-t, el sem birom képzelni. - Rugalom Tzs. (1835)

The relevant part is that the author found rugalmas already in 1829 but rugalom only in 1835. These dates are probably not correct because the tools for searching through large bodies of text were very inefficient at the time, so they often missed early uses. But still it shows that it's conceivable that rugalom was a back-formation.

At this point we can only state reliably that it's rúg+-alom+-as. This is consistent with both possibilities and refrains from taking one explanation over the other.

Now, looking at the big picture, things like this happen quite often in Hungarian because the language doesn't really work sequentially like building a new word with each suffix where you have a ((((stem)suffix)suffix)suffix) structure. Sometimes multiple suffixes get attached at the same time. The speakers still feel that they are multiple suffixes but they may not consider the intermediate form as a proper word. Later they may change their minds and start using the intermediate form, too. Now you might say they back-formed it, but they actually "knew it all along" that there is a split there, it was just not customary to use that intermediate form.

In some cases, however, multiple suffixes may may start to fuse into one mental unit and we can consider those to be "compound suffixes". We have some entries like that, e.g. -lkodik, -skodik, -sít, -edék. We are not totally systematic in this. For example at fogyatkozik we just have fogy+at+kozik and no separate -atkozik entry (but I think we should have it). I think the criteria for a compound suffix should be

• many words use the constituent parts in combination only, e.g. no "fogyat", only "fogyatkozik"
• significant change in meaning compared to what would be expected from concatenation, e.g. -hatatlan, which is actually -hat + -atlan but the meaning of -hatatlan does not follow from just the two parts as they are used individually.

Also sometimes back-formations are later treated as the base form and people actively derive the original form as a derived form from the back-formation. For example, sikér once meant gluten then sikeres was a derived adjective meaning "gluten-y" (of wheat) as a positive thing, so it's meaning changed to "successful". Then siker (success) was derived from this, and then again the -es suffix was added to it to form sikeres again, but it coincides with the earlier form, so we can't really attest this.

What I wanted to show is that (((stem)suffix)suffix)suffix) model is not always fitting as back-formations and "sleeping intermediate forms" make the situation less conceptually clear.

Suggestions:

• Keep rugalmas as rúg+alom+as, at least until we find some reference that tells more. Do the same for similar undecidable cases.
• Create entries for compound suffixes when it makes sense. I wouldn't do it for -almas (because there is no change in meaning and most words with -almas have the corresponding -alom form as well). But I'd do it with -atkozik. Qorilla (talk) 23:40, 26 October 2014 (UTC)

When several patterns of derivation form a common whole, then speakers might sometimes coin terms that skip the intermediate steps. And sometimes the intermediate steps are actually back-formations. Hungarian is definitely not the only language where this happens. I'm not really sure what should be done when an intermediate form is expected, but not attested or actually attested much later. I wonder what the other editors think of this.