A discussion in another thread led me to pose this question. The association of iĝi with intransitive verbs possibly shouldn't happen in theory, but historically it has been a part of the evolution of the language, meaning variously iĝi X-anta (with sid'), or iom post iom X-i (with ĉes', or tranvil'), and there may be other possibilities.
It may be worth asking: should it be encouraged at all? Or should a more "logical" practice prevail? What is your view?

The logic here Miland is just, does the compound with 'iĝ' make sense - has it got a real world application. If so it is fine.

I tend to think of them as slightly different.

In the case of intransitive verbs, -iĝi (to me, at least) implies an active process of "becoming" that the plain verb doesn't always have.

Now, for some verbs, there may be no difference between the two. But I think for a lot of verbs, that, perhaps subtle, difference is there.

In the case of "sidi" vs. "sidiĝi", for example - being the example from the other thread - I personally think there is a difference between "sitting", and "becoming seated".

Mi sidas. I am sitting.
Mi sidiĝas. I sit down (or, I become seated).

Note that "mi sidas" is also (maybe even more commonly) translated as "I sit", but I think my translation above more clearly demonstrates the difference I perceive between the two expressions. The one shows, or at least, can show, a rather static action, whereas the other is always active.

To me -iĝ in an intransitive verb has both a logically clear meaning and valid uses (as well as traditional acceptance), so I feel no need to discourage it. Except perhaps in specific instances where use of ek- may be more suitable.

For a verb like sidi I see a small difference in nuance between eksidi and sidiĝi. They both mean approximately the same thing: that you started sitting. But use of -iĝ gives extra emphasis to the idea of coming into a seated state, or RiotNrrd's "become seated". Since sitting (or being seated) doesn't really involve much dynamic action, I believe that in many cases this is closer to the idea people have in their minds. Possibly that would explain, at least in part, the pervasiveness of sidiĝi in the language.

I note that in Tekstaro sidiĝ' is much more popular than eksid' (811 hits v. 136). Though interestingly this isn't mirrored in Google, where all forms of eksid' are some 3 times more common.

I looked up Zamenhof's Lingvaj Respondoj and found more or less the discussion so far summarised; a correspondent wrote to Z. saying that one should not use iĝi after intransitive verbs, but rather ek before. Z. agreed, but said that prevalent usage and practicality made the implementation of usage that was logically perfect impossible, but that on the other hand in an artificial language, unlike natural languages, people always had the right to adhere to a more logical usage.
An internet version can be found here, but without the footnote in the printed edition (which also distinguishes ek from iĝi). See the section "PRI “SIDIĜI” KAJ “EKSIDI” ", about a quarter of the way through (para 13 in printed editions).

Miland:I looked up Zamenhof's Lingvaj Respondoj and found more or less the discussion so far summarised; a correspondent wrote to Z. saying that one should not use iĝi after intransitive verbs, but rather ek before. Z. agreed, but said that prevalent usage and practicality made the implementation of usage that was logically perfect impossible, but that on the other hand in an artificial language, unlike natural languages, people always had the right to adhere to a more logical usage.
An internet version can be found here, but without the footnote in the printed edition (which also distinguishes ek from iĝi). See the section "PRI “SIDIĜI” KAJ “EKSIDI” ", about a quarter of the way through (para 13 in printed editions).

thank you for the link, Miland. In the section that you mention above, it's particularly interesting to read Zamenhof's opinion on the matter of certain language forms being more commonly used, even if they are perhaps not the absolutely most logical. I like the fact that he is quite flexible on this, which seems to be a sensible approach. Zamenhof was clearly happy to let time show which form attracts the most users.

This approach also applies to the other thread we had here about 'kiu' or 'kio' in the question 'What is your name?'. I would say that 'kio' is still most widely used in this situation, which would suggest that the arguments of logic in favour of 'kiu' have not convinced enough people for it to fully catch on. Still, people can perfectly well use it, if they believe this is the best way. Ultimately, I think everyone understands both forms perfectly well. I have major difficulties believing that those in the favour of 'kiu' have the immediate reaction of answering 'My name is a word by which I am known!' when they hear 'Kio estas via nomo?' Even the most extreme proponent of absolute logic would surely answer by saying their name!

RiotNrrd:Mi sidas. I am sitting.
Mi sidiĝas. I sit down (or, I become seated).

Note that "mi sidas" is also (maybe even more commonly) translated as "I sit", but I think my translation above more clearly demonstrates the difference I perceive between the two expressions. The one shows, or at least, can show, a rather static action, whereas the other is always active.

This is absolutely true, but you're comparing 'sidiĝi' with 'sidi', rather than 'EKsidi'.

When you think of 'EKsidi', then this clearly means that you 'started sitting' and how do you do that? You sit down! So, I'm having a very hard time seeing the difference between 'eksidi' and 'sidiĝi'!! It probably just doesn't matter which one you use.

Quite right Chainy, and an extreme proponent of absolute logic should answer Kiu estas via nomo with the question Kiu el kiuj.

Looking through a list of poorly written names in a hotel register you might ask Kiu estas via nomo to determine which line someone had made an entry on. In those circumstances the question Kio estas via nomo might not be the obvious choice if the name was already known, but the problem was locating the name.

Chainy:So, I'm having a very hard time seeing the difference between 'eksidi' and 'sidiĝi'

The difference as I see it (and I admit it's only a small one) is in emphasis. I tend to think of sidiĝi as the English "get sat down". When you get sat down you of course start sitting, but the emphasis is on the change to a sitting (or seated) state rather than the commencement of action. If you wanted to use linguistic terms I would perhaps put this down to the difference between the incohoative and ingressive aspects.

I understand that sidi is a verb and that verbs denote an action, and that the idea of "becoming" is something more associated with states than actions, and thus become-sit could seem a bit wierd in comparison to things like ruĝiĝi (become-red) or rompiĝi (become-break) where the idea of state and state-change is ostensibly inherant to the meaning of the root. But there are quite a few verbs whose meaning suggests state at least as much if not more than action. The most obvious example, though probably irrelevant to what we're discussing, would be the so-called stative verbs. I don't class sidi as a stative but to me it implies state through it's basic meaning. A better example might be "boli"; consider why you will often see something like "esti en stato de gasiĝo" in it's definition. So "boliĝi" doesn't feel like least bit odd to me. I feel similarly for other intransitives like bruli, stari, sveni and odori. For more dynamic intransitives like kuri, ŝpruci, salti etc I would find "ek" or "komencis" more logical.

With regard to the lingvaj respondoj, Zamenhof appears to have been answering a question specifically about sidiĝi/eksidi so I'm not sure you can extend his comments to the practice of using -iĝ in intransitive verbs ĝenerally. Anyone got the transcript of the question that was put to him so we could verify that?

But I like Z's observation that "se formo estas efektive bona, ĝi baldaŭ trovos multajn imitantojn kaj iom post iom elpuŝos la malpli logikan". If the iĝ-intransitives ever do get "pushed out" then perhaps I'll revise my viewpont but for the time being it seems that someone finds them useful.

Here's the footnote from the printed version, as it may be of interest. It is referenced by a superscript "1" just after the first word of the passage "Vi":

Aymouler (kaj Huet) instruis, ke "iĝ" post transitivaj verb-radikoj signifas ĉiam "iĝi-ata", kaj ke post netransitivaj oni devas uzi ne "iĝ" sed la prefikson "ek". Malgraŭ la tolerema cedo de Zamenhof, la formo "pli logika" ne elpuŝis la "malpli logika". Cetere, kiel demonstris Corret, estas ia nuanco inter "iĝ", kiu signas la pason de unu stato al alia, kaj "ek", kiu signas la komencon de alia stato, aŭ ago.

I translate:
"Aymouler (and Huet) taught that "iĝ" after transitive verbal roots always means "iĝi-ata", and that after intransitives one should not use "iĝ" but the prefix "ek". Despite the tolerant concession of Zamenhof, the "more logical" form has not pushed out the "less logical". Besides, as Corret demonstrated, there is a certain nuance between "iĝ", which denotes the passage from one state to another, and "ek", which denotes the beginning of another state, or act."

There is no further information about Corret.