A technical comment in Science shows how quickly an evolutionary just-so-story can unravel (“Comment on ‘Phonemic Diversity Supports a Serial Founder Effect Model of Language Expansion from Africa,'” February 10, 2012). Quentin Atkinson, last year in the same journal, had proposed the striking and suspiciously neat theory that worldwide linguistic evidence, based on the size of the phonemic toolbox of languages around the world, showed that language had originated in western African between 50,000 and 70,000 years ago.
The seemingly commonsensical basis for this was the observation that as you fan out from the region, home to the most phonemically rich language community, phonemic inventories get progressively simpler and sparser — with the simplest and sparsest being found among Polynesians and Australians, who presumably had the least opportunity to diversify theirs. (A phoneme is the smallest unit of sound that goes to making up words.)
The problem is this turns out to be a bit of a fable, based on partial and imperfectly analyzed evidence, as Michael Cysouw and his colleagues show in hardly more than a page. The full data suggest no such simple pattern:

[W]e see the following problems with Atkinson’s findings. First, his data are coarse-grained summaries of the UCLA Phonological Segment Inventory Database (UPSID) (3) as reported in the World Atlas of Language Structures (WALS) (4). To illustrate our concerns, we used the original UPSID database (which is freely available online) together with tone data from WALS. Atkinson’s WALS-based estimates of phoneme inventory size turn out to be only imperfectly correlated with the actual number of phonemes as specified in UPSID (r = 0.60, P < 2.2 � 10?16). Specifically, his WALS-based data give unjustified weight to the number of vowels and tones at the expense of the number of consonants, strongly biasing the resulting geographic patterning toward western Africa's having large phoneme inventories (figs. S3 and S4). When the UPSID data are appropriately corrected for speaker community size and linguistic genera through a mixed-effects model, the largest phoneme inventories are actually found in North America.

Yes, that’s among Native Americans here in the Pacific Northwest and neighboring Canada.
It’s always tempting to see that one pattern in the carpet and miss all the rest. In fact, we don’t know where language originated. A linguist friend of ours, however, recommends:

The Khoisan languages of southern Africa really do have far and away the most complex sound systems on earth — lexical tones, clicks, everything. Those ingressive clicks, in fact, are found only in South Africa. To hear some of these sounds you might go here and see if you can access any of Peter Ladefoged’s sound files hosted at the UCLA Phonetics Lab.