From: Robert Blust [blust@hawaii.edu]
Date: Friday, May 10, 2019 3:47 PM
To:   R David Zorc; Robert Blust
RE:   Important correction to PPH *puruq 'island' => *pujuq

Dear David,
  Thanks very much for your fix on this one ('island'). I have to admit I should have caught this myself earlier, but somehow missed it.  *j is always a little tricky, since it is of low frequency, 
and its reflexes are sometimes phonetically quite distinct. In any case I have replaced the *r with *j. The good thing, of course, is that this now strengthens the comparison considerably, providing better evidence for connecting N + S Philippines in a unified group.
  As for *-q vs. glottal stop I am still not convinced that I should recognize a PPH *7, for the same reasons I resist the temptation to include it in the PAN or PMP inventory, all of which is discussed at some length in my online Austronesian book. It is true that *-q normally yields Agutaynen -k, and if I have counted everything correctly there are 14 instances of this currently in the ACD:
 1. *lemeq > lemek 'soft'  {lemeq~'soft~flexible'-Blust[ACD].docx}
 2. *luSeq > lok 'tears'   {luhaq~luSeq'tears'-Blust[ACD]+Zorc.docx}
 3. *dilaq > dilak 'tongue' {dilaq'tongue~lick'-Blust[ACD].docx}
 4. *lutuq > lotok 'cook'   {lutuq'cook'-Blust[ACD]+Zorc.docx}
 5. *taRuq > talok 'hide'   {taRuq'store~hide'-Blust[ACD].docx}
 6. *tulduq > toldok 'to  point' {tulduq'finger(index)_point'-Blust[ACD]+Zorc.docx}
 7. *piliq > pilik 'choose'  {piliq'choose~select'-Blust[ACD]+Zorc.docx}
 8. *tabeq > tambek 'fat'  < Bs *tambek | Otherwise SEE: {taba7~tabeq'fat~grease'-Blust[ACD]+Zorc.docx}
 9. *peceq < pantek 'hatch'  {pesaq~peceq'break(pieces)'-Blust[ACD]+Zorc.docx}
10. *taneq > tanek 'land'   {taneq'earth~soil'-Blust[ACD].docx}
11. *nanaq > nanak 'pus'    {nanaq~naNaq'pus'-Blust[ACD].docx}
12. *panaq > panak 'shoot'  {panaq'bow&arrow'-Blust[ACD]+Zorc.docx}
13. *salaq > talak 'fault'  {salaq'mistake'-Blust[ACD]+Zorc.docx}
14. *suluq > tolok 'torch'  {suluq'torch'-Blust[ACD]+Zorc.docx}

However, there are also 9 instances of *-q > zero currently in the ACD:
1. *qaq > ka 'indeed (confirmation particle)' {a~qa'indeed'[dp-confirmation]-Blust[ACD+PPH}+Zorc.docx}
2. *qiduq > kiro 'dog'      {idu7~qidu7'dog'-Blust[ACD]+Zorc.docx}
3. *langkaq > langka 'jackfruit' {langka7'jackfruit'-Blust[ACD}+Zorc.docx}
4. *daliq > ma-dali 'quick'   {dalíq'quick~quickly'-Blust[ACD]+Zorc.docx}
5. *sub(e)liq > tobli 'pass e.o. in different directions'
{subli7'pass_going_in_different_directions'-Blust[ACD+PPH].docx}
6. *tandaq > tanda 'remember, recall'  {tanda7'memory(good)-intelligent'-Blust[ACD+PPH]+Zorc}
7. *pasuq > paso 'burn skin'  {pasu7~pasuq'roast~scald'-Blust[ACD]+Zorc.docx}
8. *pujuq > poro 'island'  {puju7=púruq'island'-Blust[ACD+PPH]+Zorc.docx}
9. *tingadaq > singara 'look upward'  {tingadaq'look_upward'-Blust[ACD].docx}
  Some of the latter may well be loans from Tagalog or Bisayan sources (as paso), but this seems very unlikely for #1, and not particularly likely for several others.  For this reason I am going to use *pujuq to represent this form until I am convinced by better evidence to replace the *q with glottal stop.
  It is finals week here and I am going 'brain dead' reading, providing lengthy typed comments on, and grading term papers for our grad students, and I am not done yet, so will stop here for now.  
I'll respond to your longer message of a few days ago once I get a break.
With all good wishes,  Bob
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Date: Thu, May 9, 2019 at 12:30 PM
From: R David Zorc <dzorc1@comcast.net>
Hi again, Bob,
  I’m still going through the article (and will be for weeks, if not months). Here is one for which I have had data since 1982. Note that the data sheet from my ZDS is two pages, one for NPh supporting penult length, and the second for SPh supporting accent on the ultima. There are several examples were NPh and SPh languages collectively reflect a difference in accent. Yet another reason for my proposing a PNP node as opposed to a PSP one. Also, the Kalamianic 
languages clearly support a glottal stop as the final consonant. Unlike the Inland Manobos where the knowledge of “islands” is clearly second-hand and therefore the word is a loan in WBM), the Agutaynen live on an island and the Kalamian Tagbanwa are well aware of what an island is. Yet neither of these languages show a final –k. I consider this one of the better-established PPH reconstructions which supports the PPH node (with shift in accent occurring as they journeyed southward) and the need for an intervocalic *j (Ilokano being the odd man out) and the need for a final glottal stop.
 By the way, do you have McFarland’s Northern Philippine Linguistic Geography. I have found it a treasure-trove of exciting reconstruc-tions. There are dozens and dozens of remarkable “innovations” among the languages he treated.
Cheers!  David