tommjames:...There's a passage in PMEG...

Great! That is exactly the kind of explanation I was looking for. So, if I really needed to insert a vowel in a case such as this, I would use -e. But this is frowned upon. Thanks for the answer and citation.

Of course, by now we are way off topic

I would still like to hear from people on the original post if anyone has thoughts.

seveer:
I would still like to hear from people on the original post if anyone has thoughts.

I tend to hold geminated consonants rather than pronounce each separately. For continuants this is straightforward. For stops there is a fraction of a second of silence before completing the consonant. For the r I use a flap for the single r and a trill for the geminated r. As tempodivalse said, there is some amount of interpretation involved in Rule 9.

Zamenhof did say Esperanto should sound somewhat like Italian, and the pronunciation I go by follows Italian gemination principles.

Bruso:

seveer:
I would still like to hear from people on the original post if anyone has thoughts.

I tend to hold geminated consonants rather than pronounce each separately. For continuants this is straightforward. For stops there is a fraction of a second of silence before completing the consonant. For the r I use a flap for the single r and a trill for the geminated r. As tempodivalse said, there is some amount of interpretation involved in Rule 9.

Zamenhof did say Esperanto should sound somewhat like Italian, and the pronunciation I go by follows Italian gemination principles.

I tend to follow Japanese (which is like Italian, maybe?) in my pronunciation. It is a bit of a softer pause between the consonants than I think Japanese calls for, though. As such, I tend to geminate for consonants that Japanese frequently allows and shy away from the ones it doesn't---so, mostly unvoiced consonants.

As to whether or not this is good, prescriptive practice? I've found this thread very heartening in that it seems to indicate that I'm at least generally on course!

I mostly try to pronounce geminate consonants the same way in Esperanto that I do in English. Admittedly, they are very rare in English without occurring across word boundaries (and in fact the only minimal pair I can think of offhand is unaimed vs. unnamed), but those word boundaries are the benchmark I use. I try to pronounce geminate D as in "bad dog", geminate G as in "big gun", and so on. So, I try to pronounce the "kk" in ekkrii the same way I pronounce the "ck c" in black cat.

As for whether or not they should be avoided...I don't know. I tend to think an international language should be as easy to pronounce as possible for the widest possible slice of the human population, but I also tend to think that if that pronunciation were part of the goal of Esperanto, it wouldn't have nearly as many consonant clusters as it does, and would probably look more like Hawai'ian than Italian. Sadly, I am not enough of a polyglot to claim any real authority on what native speakers of other languages do or don't find difficult to pronounce. With that in mind, I usually try to just form words as necessary, without tending toward or away from geminate consonants. My English roots would tend to avoid them, so I try not to let my native tongue influence my Esperanto any more than it has to.

Breto:I also tend to think that if that pronunciation were part of the goal of Esperanto, it wouldn't have nearly as many consonant clusters as it does, and would probably look more like Hawai'ian than Italian.

Esperanto's vocabulary was planned to be recognizable to educated Europeans (or non-European speakers of European languages). If you've seen what happens to words borrowed into Hawai'ian, they're not very recognizable. Kikiki=ticket. Welaweka=velvet. Palaoa=flour.

Some critics like to attack Esperanto's phoneme choice, but it is really essential if you want to import Greek and Latin-based lexicon, which is (whether you like it or not) widely in use around the world, ranging from normal layman terms like orkestro to technical or scientific vocabulary - ontologio, aksiologio. Removing phonemes would make those roots unrecognisable, since you would have to replace them with other phonemes (like in Hawaiian, and to a much smaller extent Italian, which has an odd aversion to consonants).

Bruso:

Breto:I also tend to think that if that pronunciation were part of the goal of Esperanto, it wouldn't have nearly as many consonant clusters as it does, and would probably look more like Hawai'ian than Italian.

Esperanto's vocabulary was planned to be recognizable to educated Europeans (or non-European speakers of European languages). If you've seen what happens to words borrowed into Hawai'ian, they're not very recognizable. Kikiki=ticket. Welaweka=velvet. Palaoa=flour.

That's what I was getting at. Esperanto values recognizability over ease of pronunciation. I'm not saying that's good or bad, just that it is the choice that was made. And as Tempodivalse has pointed out, it gets more and more difficult to recognizably construct Greco-Roman technical language the fewer distinct phonemes you have to play with.

@Tempodivalse: That said, I'm not sure that "aversion to consonants" is quite as odd as you make it out to be. I don't have specific numbers in front of me, but I'm pretty sure there are more than a few languages worldwide that are much more limited on which consonants can come together in a word without an intervening vowel than English is, and even English is more restrictive than Esperanto.

Breto:@Tempodivalse: That said, I'm not sure that "aversion to consonants" is quite as odd as you make it out to be. I don't have specific numbers in front of me, but I'm pretty sure there are more than a few languages worldwide that are much more limited on which consonants can come together in a word without an intervening vowel than English is, and even English is more restrictive than Esperanto.

Japanese would be on the top 3. It hates consonants with passion, I can say. Consonant clusters are almost non-existent, only sh, k, s and n can go together with another consonant, of course, only -n and -s can end a word. Sakusen, kashite, desu are pronounced saksen, kashte and des in normal speech. Japanese is a syllabic language, even its writing system, which firstly uses 2 syllabaries (Hiragana and katakana), is made like that, syllable after syllable. Though, Japanese does use gemination a lot, I’d say much more than Italian. Japanese has few sounds in general* and a vowel is always part of a word, because of this pretty much any language on Earth is hard to pronounce for JP speakers, especially English. I can’t imagine how hellish it must be for a JP to learn Polish, which is the yin of Hawaiian, Polish abuses consonants, too many consonant clusters, too.

*This at the same time poses a huge disadvantage; Japanese has an absurd amount of homophones due to its limited phonological repertoire and lack of consonant clusters. It is much more context dependent than many other languages.

*This at the same time poses a huge disadvantage; Japanese has an absurd amount of homophones due to its limited phonological repertoire and lack of consonant clusters. It is much more context dependent than many other languages.

Mandarin has the same problem - and even the presence of tones is unable to disambiguate in a lot of cases.

My 1st language is Russian, which is probably the most cluster-friendly of the major world languages, with the possible exception of Polish (e.g., kontrprimer, vsprygnut' ...).

As a consequence, I barely notice Esperanto's consonants (eksprezidento), and I find it mildly funny that Italian almost never has two pronounced consonants in a row, unless one of them is s or r - so consonants mysteriously disappear: from "example" you get esempio.

Tempodivalse:

*This at the same time poses a huge disadvantage; Japanese has an absurd amount of homophones due to its limited phonological repertoire and lack of consonant clusters. It is much more context dependent than many other languages.

Mandarin has the same problem - and even the presence of tones is unable to disambiguate in a lot of cases.

My 1st language is Russian, which is probably the most cluster-friendly of the major world languages, with the possible exception of Polish (e.g., kontrprimer, vsprygnut' ...).

Certainly Mandarin suffers that terrible fate, too, but it isn't as bad as Japanese thanks to the tones even though they themselves don't always help, as you said. Rather, the problem of Mandarin is the completely lack of morphology as opposed to the agglutinative nature of Japanese with its heavy morphology. Languages like those two are way easier to read than to listen to. Kanji/hanzi are a bless, they’re crystal clear in many ways, they’re the best tool against homophones.

Russian is fairly easy to pronounce, actually xD. Fortunately it doesn’t abuse consonants like Polish does. I put Russian on pair with Spanish when it comes to clusters, though Russian may have a bit more clusters.