Leo adds to Scott:
Your wife probably has full command of a variety of phonemes that exist in Chinese languages other than Mandarin. You may not have noticed this, and she may not even be aware of all her skills. She just uses them.

-----Original Message-----
From: International Auxiliary Languages [mailto:[log in to unmask]] On Behalf Of Scott Raney
Sent: Tuesday, July 29, 2014 6:37 PM
To: [log in to unmask]
Subject: Re: World language phonemes (was Re: Source languages of Pandunia)

On Tue, Jul 29, 2014 at 10:24 AM, Risto Kupsala <[log in to unmask]> wrote:

> As I explained earlier by that "bin-spin-pin" example, it happens to 
> be so that English and Mandarin are somewhat compatible.
> In words 'bing' and 'ping'
> 1a. English 'b' is voiced and unaspirated.
> 1b. Pinyin 'b' is unvoiced and unaspirated. (Notice the difference?) 
> 2a. English 'p' is unvoiced and aspirated.
> 2b. Pinyin 'p' is likewise unvoiced and aspirated. (No difference 
> here.)
> The difference between English and Mandarin is so small that it can be 
> ignored. However, there are other languages to consider, for example 
> French, which has only voicing distinction:
> 1c. French 'b' is voiced and unaspirated.
> 2c. French 'p' is unvoiced and unaspirated.
> The result:
> The French speaker says 'bing', the Mandarin speaker hears it as 'bing'.
> The French speaker says 'ping', the Mandarin speaker hears it as 'bing'
> again. In other words, Mandarin speakers don't hear voicing distinction.
> They would only hear aspiration distinction, but that doesn't exist in 
> French. The phonetic systems are incompatible.
> Get it?

Got it, thanks. Especially for the confirmation that both "b" and "p"
(as pinyin) are unvoiced in Mandarin.  Unfortunately this whole thing has a kind of Twilight Zone feel to it to me, and also to my wife.
Although we both had a vague awareness of the difference between English and French "p", we certainly couldn't articulate it (pun intended ;-)  She majored in French in college (in Taiwan) and doesn't recall any problems with her French-speaking professors having trouble with things like bing/ping in Mandarin (my experience with French is much more limited: 2 years in high school, a semester in college, and the Michel Thomas course).  She's also sure that she's pronouncing "Bing" (English for the Microsoft search agent) the same as "bing"
(Mandarin for "ice"), and it sounds the same to me too.

All of which I take as evidence that although the linguistic analysis you have provided may be relevant if someone had the goal of speaking a language with no hint of an accent, it's simply not appropriate to build these kinds of rules in "a priori" if the goal is to create an auxlang that is merely intelligible to speakers with different native languages.  It also validates my proposal that the *only* reasonable way to proceed here is to actually do the experiments with real people and *see* what they can or can't adequately produce and/or understand.
I'm sure I have a terrible English accent when I speak French, but had little or no trouble with people understanding me in the months we spent in Martinique (assuming I could come up with the vocabulary, which usually gave me trouble at some point during an interaction).
Giving up a b/p distinction because it *might* give *some* combinations of people *some* problems in *some* word contexts is definitely a case of overcompensating.  We don't need or indeed even
*want* to design a language that everyone can speak without an accent.
Like Steve recommends with Inlis, making the language sound funny/weird/exotic to/from *everyone* may actually be the key to making it "cool", and would certainly be a great equalizer if nothing else.

I am starting to believe that my generation system needs to have a more robust speaking/listening component than I originally planned, however.  The great thing about that is that will will provide us with the empirical evidence we need to make the right choices.  Along with judging the intelligibility and "coolness" of a proposal (grammar, vocabulary, etc.), maybe have the judges speak the sentence too, then listen to a couple of samples from judges with different native languages and rate those on intelligibility.  Do this for a b/p distinction with French speakers and Mandarin judges and we'll know for sure whether we're dealing with a real problem with intelligibility, or just another case of people speaking with an accent.  Same for r/l (in various positions in words) to and from Japanese speakers.  Maybe some particular combination is a problem, maybe it's not.  Better to know for sure before we just start making overgeneralized rules that toss out valuable phonemic resources.

> --Risto