tommjames:I don't think the assumptions of a beginner necessarily show what's "natural". I imagine a lot of beginners will just look at korekta and assume it means correct because it looks similar.

That's true I suppose!

tommjames:They're not yet accustomed to the fact the root korekt- is verbal (evident from usage). If they were, then the meaning of 'corrective', which actually is the more natural meaning, would probably have been more likely to occur to them.

But it's already also clear that "korekt-" is used adjectivally for "korekta" (correct) from usage as well. It's not just inexperienced beginners using it this way. So what you're saying here still means that "corrective" is more natural only once they've been told that its root class is unchangingly verbal.

However, I am starting to feel that "korekt-" is actually a bad example for the purpose of showing why it's more natural to think of base forms instead of root classes anyway. And that "correct" is a translation worth avoiding. Not because of a verbal root class, but because this wordbuilding strategy doesn't seem to be as productive; it only sorta works on some verbal roots, e.g. komplika, konfuza, suspekta. And since it's not consistent for most verbs, it's probably worth favouring more regular approaches.

I feel like even the base form theory doesn't predict that korekta = correct. And that this usage is due to a different irregular pragmatic reason.

A better example we've heard is "senpova". It illustrates clear circumstances under which root class theory leads you astray very unnecessarily:

[list=]
"senpova" means "powerless" (sen-povo-a)
Under root class theory, "pov-" is inherently verbal. So "senpova" must mean "sen-povi-a", which it doesn't. So under root class theory "senpova" (and any construction like it, is an exception).
Under base form theory, the base form of "pov-" is "povi", so we assume a verbal interpretation in the absence of any other data (gaining all advantages of root class theory). However the "sen" suggests that "pova" is not derived from the base form, but the o-form. So base form theory explains logically what would otherwise be an exception under root class theory.[/list]In this way, I see base form theory as a more flexible refinement of root class theory. And of course there are still irregularities that it doesn't explain (like korekta). But it has more advantages than root class theory.

sudanglo:
Kiel eltrovi la propran signifon de radiko?

Ofte la propra signifo de radiko estas pli-malpli evidenta per si mem, sed por certeco necesas zorga komparo de la diversaj uzoj de la radiko kun diversaj finaĵoj, en diversaj kunmetaĵoj k.t.p. Decida estas la O-formo de radiko, ĉar la signifo de la O-formo estas ĉiam identa al la signifo de la radiko mem.

How does this tells us that manĝ is a verbal root if the most common use of manĝo is for a meal.

How does this illuminate that we use foro/flori (a flower, to flower) but only rozo and not rozi, or shed light on the most common use of fumi (to smoke cigarettes/pipe), or clarify the pair lumo/lumi.

Indeed, that idea does seem to fall down in cases where you really need it!

This comes a little late, as the database crashed yesterday when I was typing this, but...

EldanarLambetur:I'm not sure why everyone's having such a hard time agreeing with Sudanglo.

I won’t presume to speak for everyone else, but my own beef with this forum’s gentleman from Southern England in this lilliputian squabble comes down to this:

• Esperanto has effectively had a Received Grammar for a long time now. It may not be the best grammar — it may even be the worst, for all that it matters — and every seasoned Esperantist likely has issues with this or that aspect of the language, and their own way of interpreting some of the trickier grammar bits. But to deny that the Received Grammar exists, or that it never existed until recently, or that it exists, but only in the fever dreams of linguistic theorists and not on the printed page of every beginner’s handbook since the Unua Libro, or, most irritating of all, that one’s own personal interpretation of the grammar is in fact the de facto Received Grammar, is intellectually dishonest and confusing to the novice.

I don’t mean to suggest that Sudanglo or anyone else shouldn’t feel free to chime in with their own takes on the language. They absolutely should. But they should present them as exactly that. “To answer your question, earnest beginner, the prevailing rule is thus and so, though personally I find it easier to do such and such.”

• It appears not to be enough for Sudanglo to evangelize his distaste for prevailing root theory across many threads whether it’s the thread’s topic or not, turning a local debate into a regional conflict; he must also demonize grammarians as a class with all the vitriolic excess of a newly converted Marxist — and without any sense of irony that, as a man who daily holds forth on what one should or shouldn’t do with one’s grammar, he, too, is part of this parasitic class. (Which may add entertaining flavor for Eldanar and perhaps others, but, combined with the above bits, adds to the general obnoxiousness.)

Mind you, when he’s not on the warpath, he’s made some genuinely helpful posts. And “je la barbo de Zamenhof”, from a recent post of his — would that I had occasion to work that into conversation in private life.

Oh, and since it keeps coming up, if korekti = senerarigi, then korekta would logically have to mean senerariga (as it does in the PIV), not senerara. But as TommJames (I think) reminded me in this thread or some other, Esperanto’s endings are occasionally stretched to mean all sorts of things that defy the stated rules ... which makes the tenacity of this debate a wee silly.

Eldanar:So what you're saying here still means that "corrective" is more natural only once they've been told that its root class is unchangingly verbal.

That's not what I'm saying, and I don't think what I'm saying implies that either. The reason I say corrective is more natural, or even more "proper", is that in general (across the language as a whole, not just in respect of korekt-) when you take a root with an aga-signifo and turn it into an adjective you usually get certain type of meaning resultant from that, and the meaning is closely associated with the meaning of the verb. So if you a have a verb meaning "ĝustigi", then the adjectival form that fits best within that scheme is "ĝustiga" (corrective).

Yes, korekta is used for ĝusta, in the same irregular way as the other examples I gave. That doesn't necessarily make them natural.

But we can define "natural" in different ways. For some people this will just be some subjective thing, based in part on how similarly the word functions to the 'equivalent' word in their native language. Other people may prefer to compare the word with other similarly formed words in the language and see if it fits the scheme. I consider the latter approach to be preferable because it leads to more consistent usage, and consistency is A Good Thing ™.

Eldanar:I am starting to feel that "korekt-" is actually a bad example for the purpose of showing why it's more natural to think of base forms instead of root classes anyway. [..] it's probably worth favouring more regular approaches

Agreed. This is why I think the evitinda designation for korekta can be justified. Nothing to do with theory, but how logical the meaning is in the wider context of word-building in Esperanto.

Eldanar:Under root class theory, "pov-" is inherently verbal. So "senpova" must mean "sen-povi-a"

No, that's just not correct. The theory explains that flankelementoj such as prepositions may change the part of speech that the root stands for in a compound. This is not the same thing as "changing a root's class", and it doesn't invalidate the idea that roots have inherent class.

Eldanar:Under base form theory, the base form of "pov-" is "povi", so we assume a verbal interpretation in the absence of any other data (gaining all advantages of root class theory). However the "sen" suggests that "pova" is not derived from the base form, but the o-form. So base form theory explains logically what would otherwise be an exception under root class theory.

This cannot possibly be an exception under the root class theory because the root class theory states exactly this in its vortefikaj reguloj. Where you say "sen suggests that povo is the base form", root-classes says that "sen effects what part of speech the root stands for in a compound". It's the same thing, just different language. (What sudanglo claims to be a convoluted superstructure of Latinate obscurity peddled by pontificating naked emperors).

I feel like even the base form theory doesn't predict that korekta = correct. And that this usage is due to a different irregular pragmatic reason.

Right. I wouldn't argue that vort-derivo does account for the meanings of korekti and korekta.

I think that what happens in Esperanto is that usually a word is adopted in a particular part of speech meaning, so as a noun adjective, verb etc, and then derived words preponderantly relate to that meaning.

This can be arbitrary as in the classic example of broso and kombi - it is not immediately obvious why Esperanto should not have adopted brosi and kombi, or alternatively broso and kombo (for the implements).

Or it can seem to be natural - legi and kuri are not surprising adopted as the primary forms. There something inherently verbal in the meanings of the ideas. Similarly Domo and papero are primary because of the thingy-ness of the ideas.

But sometimes Esperanto will accept an idea in both the noun and verbal meanings present in the donor language(s) (planto and planti, fumo and fumi, lumo and lumi). Or in the case of korekt both the verb and adjectival meaning (ie to correct and correct.

When there is more than one 'base' form, then context may be needed to interpret derivations.

So mi sendis la tekston por korektigo could mean to make it correct or to have someone correct it.

Yes, korekta is used for ĝusta

A more sensitive use would reflect the difference between korekta and ĝusta. The two are not exactly interchangeable

Korekta relates to conformity with the rules (grammatical or social).

La princo kondutis korekte kiel ĝentlemano. Lia konduto estis absolute korekta.

En tiu kazo la ĝusta respondo ne estis taksita korekta.

Oni ne ĝustigas tekston. Oni korektas ĝin.

Ĝuste (ne korekte) kiam mi akiris sufican kuraĝon por inviti ŝin danci, ŝi malaperis.

The theory explains that flankelementoj such as prepositions may change the part of speech that the root stands for in a compound. This is not the same thing as "changing a root's class"

You mean like if a man wears womens clothing he is still a man.

I see it now, the roots don't change their class they disguise themselves.

sudanglo:This can be arbitrary as in the classic example of broso and kombi - it is not immediately obvious why Esperanto should not have adopted brosi and kombi, or alternatively broso and kombo (for the implements).

Russian and Polish have a word for "comb" (česat'/czesać), but a substantive for "brush" (ščetka/szczotka).
Sometimes it's that easy.

BTW: Have a look at the Universala Vortaro: Esperanto roots are translated there by words in the 5 languages, always of the same class.

----

I think a part of the confusion of the root class comes from the fact that while some roots are prototypically things, actions or qualities (ŝton/, far/, grand/), some other could equally well belong the one or the other, e.g. instruments (as objects or the action done by the object), cf. this paper by Ken Miner.

sudanglo:A more sensitive use would reflect the difference between korekta and ĝusta. The two are not exactly interchangeable.

Perhaps not interchangeable in absolutely every case, but almost. I've seen korekta used in the social/political sense before, but when used for a meaning other than "corrective" it seems to me to be usually just another way of saying "ĝusta". Not sure I see how this relates to root-classes though.

sudanglo:inherently verbal

thingy-ness of the ideas.

Just another way of saying a meaning that allows the root to belong to a certain class.

Thank you Kirilo for the link to the Miner paper. It provides some nice quotes.

Ofte oni parolas pri derivado de du formoj el tria (baza) formo, ekzemple de 'bono' kaj 'boni' de la bazforma 'bona'. Sed la afero ne estas ĉiam tiel simpla. Ofte ni trovas, ke (malgraŭ ofta opinio, eĉ dogmo, inter la E-istaro) radiko fakte ne havas konstantan signifon

Ni devas atenti interesan kaj evidentan fakton: vortaroj, inkluzive E-vortarojn, ne difinas radikojn, sed vortojn

Se kaj 'fido' kaj 'fidi' estas leksikaĵoj, kio pri 'fida'? Evidente, 'fida' povas esti derivaĵo aŭ de 'fido' aŭ de 'fidi'

Resume: kiam en triopo X-o, X-a, X-i, unu el la formoj estas prototipa por sia kategorio (estas substantivo indikanta konkretan aĵon, adjektivo modifanta karaktere, aŭ verbo predikanta konkretan procezhavan agon), tiu formo estas la bazformo, kaj la aliaj deriviĝas de ĝi per ŝanĝo de la finaĵo.

Se neniu formo estas prototipa por sia kategorio, la uzado (statistika plejofteco) determinas la bazformon. Eblas, ke pli ol unu bazformo ekzistas

And that folks is precisely the case with korekti and korekta where statistically the usage is even divided between verbal and adjectival usage.

Search the Tekstaro with korekt\VF and korekt\AF, if you don't believe me.