sudanglo:Semantically the raw data, so to speak, of the language is the use of words, not roots, which in contrast to words only have a theoretical meaning, not one you can look up in the dictionary (if they are of the type that don't need a finaĵo).

Have a look at how Zamenhof distinguished roots from words, e.g. dot' vs. dot'o.

This topic has been discussed over and over again, as long as you don't belong to a school of "anything goes" (which leads nowhere, pun intended), the root class approach is as evident as can be.

sudanglo:If blu- always means blua, this presents a problem

No! blu- does not always mean blua! The root blu- is inherently adjectival, but it doesn't mean blua. In the link I gave uses the term "kategorio de la radiko" to describe what I'm talking about. More below:

tommjames:Yes, but of course there is no problem because the root class theory does not say that blu- always means blua. In fact the theory as established by the Akademio states words to the effect of your "neat sidestep", i.e that in some words blu- could stand for bluo. Far from being "convoluted" this is explained simply in terms of a preposition (such as "sen" ) having a nounifying effect on the root. Since prepositions usually come before nouns, it's easy to see why that would be the case.

You seem to be basing your argument on the premise that the concept of roots having class is undermined by the fact other vortelementoj in a compound can affect what part of speech the root logically stands for.

I agree completely with this interpretation of root class theory. I also want to point out that the theory of la Akademio uses two different terms that may be helpful to think with: the idea of an inherent category for a root is called the "kategorio de la radiko". The idea of the semantic usage of a root in a particular situation as affected by roots to the left or to the right is called the "vortkaraktero" of the element (root or compound). That word is only used in section XIV, which has the words tommjames is talking about. Sudanglo, I think you are resisting the idea of "kategorio de radiko" because of the existence of "vortkarakteroj", which are often the same, but sometimes different. "Vortkarakteroj" are significantly more intuitive to think in, of course, but they can be derived accurately from la Akademia root theory.

Reproduced below:

XIV. NI KONSTATAS, ke ĉe duobla flankelemento ne nur la dekstra elemento povas difini la vortkarakteron de la maldekstra elemento (ekz‑e: skribmaŝina = skribo‑maŝino‑a), sed ankaŭ la maldekstra elemento povas efiki al la dekstra, difinante ties vortkarakteron. Adjektiva aŭ prepozicia maldekstra elemento efikas substantivige: grandkuraĝa = (granda kuraĝo)‑a = karakterizita de granda kuraĝo; palblua = (pala bluo)‑a; helverda = (hela verdo)‑a; senforta = (sen forto)‑a; senutila = (sen utilo)‑a; senpova = (sen povo)‑a; senduba = (sen dubo)‑a. Verba prefikso kiel maldekstra elemento efikas verbige: reformo = re‑formi‑o; ekmartelo = ek‑marteli‑o. Tio estas la “inversa vortefiko”.

Also I think these are cool questions of word semantics. These are my answers of what sounds right to my ear:

sudanglo:What is bicikla ferio? (when bicycles go on holiday, or a cycling holiday?) What is a bicikla sonorilo?

bicikla ferio means a cycling holiday, and though I'd probably translate into English biciklferio the same way, the latter has a slightly different meaning. Viewing it more as something with bicycles as a necessary part of the concept? "When bicycles go on holiday" would be something like "ferio por bicikloj" or "porbicikla ferio".

sudanglo:What is a vestejo? A garderobe, or a dressing room, or both?

A vestejo at first sounds like a dressing room. A wardrobe/garderobe would be vestaĵujo (I wouldn't even use -ej there).

sudanglo:What does korekta mean?

Korekta means correctional, but that might be in opposition to prava and ĝusta, which together cover the same sort of space as English "correct" (adj). As a side note, I tend to use pravigi and ĝustigi instead of korekti, though I guess "korekti" has a very small implication that the original was necessarily malprava or malĝusta, whereas "pravigi" and "ĝustigi" don't have that need.

sudanglo:Mi rimarkis ŝian ne-mastron de la akuzativo (valid or not?)

I'd never say that, and I'd probably look at the speaker somewhat strangely. I'd say "nemastrecon", and possibly "nemastradon". The former means to me that she generally doesn't use the accusative correctly all the time, whereas "nemastrado" means she didn't use it right just then.

What are others' thoughts on this?

When do you use "pravigi"?

Since "prava" means "having the correct opinion", and "pravi" means "to have he correct opinion", pravigi usually means something to the effect of "to justify oneself/someone's self", it often has something to do with rationalizing for an action etc. (or occasionally simply proving that one was correct in the first place). I don't really see "prava" as sharing the same space at all as "korekti" and "ĝusta", because its meaning is very specific. For example, I was confused by the part where you say you'd use 'pravigi' or 'ĝustigi' instead of 'korekti'. I can understand it perfectly for "ĝustigi", but for "pravigi", I can't really see a clear link between that and "korekti".

That roots have inherent leaning to a particular part of speech is clearly evident in the derivational morphology of a vast and diverse array of words and compounds in use.

That's fine, Tom, but is just not the same as saying that a root has an inherent grammatical class - which description seems still used in explanation.

PMEG for example talks about derivations of adjectives from roots with aga signifo.

And the John Wells dictionary talks about verb/adjective/noun roots in its preface on the grammar. It is important to note the exact meaning of the root. (He then muddies the waters by continuing with 'Profesoro (sic) is a noun root'. No it isn't 'profesoro' is a word.)

And the latest PMEG now considers 'korekta' (conforming to the rules) to be an error, whilst at the same time recognizing the frequency of use in this sense.

In fact, it is coming across this shift in position in PMEG that provoked me to revisit the topic.

I want a theory under which korekta can be korekta (not just corrective). Theory has to fit the usage - not vice versa.

You seem to be basing your argument on the premise that the concept of roots having class is undermined by the fact other vortelementoj in a compound can affect what part of speech the root logically stands for.

Yes, this 'undermining' demonstrates that roots don't always have a unique inherent class.

It is sloppy talk to say that a particular root is a verbal/noun/adjective root. It is the meaning of words that directs the meanings of compounds where the word minus the finaĵo (ie root) is a component.

Profesori is derived from profesoro - instruisto from instrui.

The basic facts of the language are the meanings of words as demonstrated by how they are actually used.

Well yes, since the root-class theory has near universal acceptance and Academy approval it doesn't surprise me to see a number of authorities make use of it's terms when explaining how the language works.

By the way, to know for sure you'd need to ask Bertilo but it seems to me PMEG's U-turn on "korekta" is based on concerns over clarity, not a desire to make Esperanto "fit into the root-class theory" (for example when seeing korekta used to mean "correct" a speaker may erroneously start using "korektigi", as in granda → grandigi, bela → beligi, etc). To me that represents a shift in opinion of the kind that could have happened with or without root-classes, so it doesn't seem to say much about the validity of the theory.

As for Wells, I'm pretty sure he understands the difference between a word and a root. I don't see how an accidental sub-optimal use of words reveals anything important.

sudanglo:I want a theory under which korekta can be korekta (not just corrective). Theory has to fit the usage - not vice versa.

Non sequitur. The desire to have korekta limited to "corrective" is not necessarily anything about conforming to some theory. Rather it is about conforming to the norms of wordbuilding logic and making Esperanto more systematic and consistent with itself.

Anyway, you have every right to demand a theory that takes all the exceptions and edge cases into account in some new all-harmonious paradigm, but you have not yet shown how or why such a thing is desirable, useful, or even possible. You have simply taken these things as givens, and this continues to undermine your argument.

erinja:Since "prava" means "having the correct opinion",

It usually only means that when applied to a person, though. You can have "pravan ideon", which would be an idea that is right/correct in your view (as opposed to "ĝustan ideon" which has a more objective sense. Same with "pravi": "Mi pravas" is "I have the correct opinion", but "Via ideo pravas" would be "Your idea is right in my (his/hers/someone's) opinion."

erinja:I can understand it perfectly for "ĝustigi", but for "pravigi", I can't really see a clear link between that and "korekti".

I'm sorry! I should have been more clear! I see "prava" and "ĝusta" as sharing the same space as the English adjective "correct", not Esperanto "korekt-". As for korekti/pravigi/ĝustigi, I just tend to use pravigi (justify, cause something to be right by trying to change the way someone views "right") and ĝustigi (cause something to be right by changing the thing itself to conform to "right"), not korekti (completely replace something with what is "right") in speech; I don't view them as sharing the same meaning. That was what I was trying to say.

All this lively discussion puts me in mind of an equally lively discussion conducted recently (actually still going on) in an Interlingua forum -- but in that case conducted entirely in Interlingua, not in English (or other "natural" language)! -- about a somewhat fine point of (Interlingua) language theory, just as this thread is about a somewhat (to me) fine point of (Esperanto) language theory. I consider it an indication in both instances of real life in each language. (Full disclosure: I have participated on one side of that debate, even more so than I have made a few comments in this E-o thread.)

The desire to have korekta limited to "corrective" is not necessarily anything about conforming to some theory. Rather it is about conforming to the norms of wordbuilding logic

Really, Tom? Doesn't limiting korekta to corrective entirely follow from the view that korekt is a verbal root and therefore korekta must mean corrective.

On the other hand if you consider the meanings of words to be the primary source for derivation, then the meaning of korekta (as established from usage) can be comfortably what it usually means and occasionally (as context makes clear) also corrective when it is seen as derived from korekti.

The potential ambiguity of korektiga (is it derived from korekti or korekta) seems to me a small price to pay for the usefulness of korekta (meaning conforming to the rules) for which meaning ĝusta is not always an equivalent.

Nobody seems to have a problem with manĝo (normally a meal, something eaten at prescribed times which may consist of several courses), but also theoretically an act of eating if manĝ is a verbal root). Why victimise korekta? Doesn't manĝo also not conform to the norms of 'wordbuilding logic' under the theory of grammatical class of root?

Why panic about potential ambiguity in korektigi (make correct, get corrected) and yet raise no eyebrow over the potential ambiguity in manĝigi la ĉevalon.