Western Malayo Polynesian Languages Micr

3
Western Malayo-Polynesian Languages

Microgroups of Mindanao
Review of Phonological, Lexical and Morphological data
Term paper
by
JEAN-FRANÇOIS DELMER
presented to Dr. Irma Peneyra

Linguistics Department
Lingg385 – Ang Pag-gugrupo ng mga Wika sa Pilipinas
January 9, 2022 – Release 1.0
+
Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 i
Contents
0 Context ................................................................................................................... 1
1 Introduction ............................................................................................................ 1
1.1 Purpose and content ......................................................................................... 1
1.2 Current classification of the languages of Mindanao ........................................ 3
1.3 Methodology: two kinds of subgroups.............................................................. 7
1.3.1 AN wave theory: Ross (1994) on "linkages" .............................................. 8
1.3.2 From lexicostatistics to the ACD to... lexicostatistics .............................. 10
1.3.2.1 Lexicostatistics - and its application to the Philippines...................... 11
1.3.2.2 Comparative word lists and dictionaries............................................ 12
1.3.2.3 Bayesian Phylolinguistics ................................................................. 14
1.3.3 The "resurrection of proto-Philippine" debate .......................................... 16
1.3.4 Blust on the importance of phonological evidence ................................... 18
2 Phonological innovations ...................................................................................... 19
2.1 Phonological innovations: literature review .................................................... 19
2.2 Phonological innovations across and within PH ............................................. 24
2.2.1 The RGH Law ......................................................................................... 24
2.2.2 The merger of *d and *z.......................................................................... 26
2.2.2.1 "Pro" arguments ............................................................................... 27
2.2.2.2 "Con" arguments .............................................................................. 27
2.2.2.3 Discussion and conclusion on the *d/*z merger ................................ 28
2.2.3 Nasal clusters .......................................................................................... 29
2.2.3.1 Experiment 1: Paz and the nasal clusters ........................................... 29
2.2.3.2 "Con" arguments .............................................................................. 30
2.2.3.3 Discussion and Conclusion of the -N- criterion ................................. 30
2.2.4 Conclusion on innovations from PMP in the Philippines ......................... 31
2.3 Focus on Mindanao ........................................................................................ 31
2.3.1 Reflexes of *R (the RGH Law) ............................................................... 32
2.3.2 The *d/z merger ...................................................................................... 33
2.3.3 The -N- prenasalized medial clusters ....................................................... 33
3 Lexical innovations............................................................................................... 34
3.1 Literature review ............................................................................................ 34
3.2 Experiment 2 - more cohesive pairings........................................................... 36
3.2.1 Revision of Ross (2020)'s calculations..................................................... 37
Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 i

3.3 Top groups clustering with South Mindanao microgroups .............................. 38
3.3.1 Gorontalic (GCP) .................................................................................... 39
3.3.2 Bilic ........................................................................................................ 40
3.3.3 Sangiric ................................................................................................... 41
3.3.4 Minahasan ............................................................................................... 42
3.3.5 Conclusion .............................................................................................. 42
4 Morphological innovations ................................................................................... 42
4.1 West-Coast Bajaw (Greater Barito / Sama-Bajaw / -) ..................................... 44
4.2 Southern Subanen (GCP / Subanon / Eastern / -) ............................................ 45
4.3 Maranao (GCP / Danao / Maranao-Iranon / -) ................................................ 46
4.4 Mongondow (GCP / Gorontalo-Mongondow / Mongondowic / -) .................. 47
4.5 Sangiric and Bilic missing... ........................................................................... 48
4.5.1 Bilic ........................................................................................................ 48
4.5.2 Sangiric ................................................................................................... 48
4.6 Conclusion on the "verbal morphology" topic ................................................ 49
5 Conclusion............................................................................................................ 50
List of Abbreviations and Symbols .......................................................................... 51
Bibliography............................................................................................................ 53
Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 ii

Tables
Table 1 - Philippines language microgroups............................................................... 3
Table 2 - non-Philippine Malayo-Polynesian languages subgrouping ......................... 5
Table 3 Regions where Philippine MP languages are spoken... .................................. 5
Table 4 Regions where Greater Central Philippine Languages are spoken... ............... 5
Table 5 Regions where Central Philippine languages are spoken ............................... 6
Table 6 Malayo-Polynesian languages microgroups spoken in Mindanao .................. 6
Table 7 Mindanao micro-groups - main references .................................................... 7
Table 8 - Top 10 most cohesive pairs (Ross 2020: 368) ........................................... 37
Table 9 - Top 10 most cohesive pairs (JFD's own computation) ............................... 38
Table 10 - Decreasing cohesiveness of pairs involving Gorontalic ........................... 39
Table 11 Decreasing cohesiveness of pairs involving Bilic ...................................... 40
Table 12 Decreasing cohesiveness of pairs involving Sangiric ................................. 41
Table 13 Decreasing cohesiveness of pairs involving Minahasan ............................. 42
Table 14 Classification of Sobanon languages in Ethnologue ................................... 45
Table 15 Classification of Mongondowic languages in Ethnologue.......................... 47
Table 16 Classification of Bilic languages in Ethnologue ......................................... 48
Table 17 Classification of Sangiric languages in Ethnologue ................................... 49
Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 1

Figures
Figure 1 Austronesian Family Tree by Blust (2009:2) ................................................ 9
Figure 2 Bayesian Tree, Philippine portion (Greenhill & Gray 2009:12) .................. 15
Figure 3 - RGH Law: reflexes by source according to Conant (1911)....................... 26
Figure 4 - Tagalog focus x aspect paradigms by Reid (1992:66) .............................. 43
Figure 5 West Coast Bajaw voice affixes by Miller (2007:64) ................................. 44
Figure 6 West Coast Bajaw intransitive affixes by Miller (2007:65) ........................ 45
Figure 7 *Mag-/*Pag- in Southern Subanen by Lobel (2013:159) ............................ 46
Figure 8 Central Subanen Verb Morphology by Lobel (2013:159) ........................... 46
Figure 9 OF verb conjugations in GCP for Lobel (2013:298) ................................... 47
Figure 10 Focus and Aspect in Mongondow for Lobel (2013:157) ........................... 48

0 Context
This end of term paper groups the responses to three of the requirements for our
Lingg385 Class:
 Req.1, "Research paper on inflectional patterns of aspect for actor and
undergoer voice affixes of 3 languages" - please refer to Chapter 4;
 Req.3, "Reports on relationships of languages based on (...) Mindanao
including Sulu" - please refer to Chapter 2 (Phonological) and 3 (Lexical)
 Req.4, "evaluation of the weight of evidentiality of witnesses of Proto-
Philippines by Blust (with citations ofc) or otherwise by Reid et al." - this is
discussed in Section 3 of Chapter 1.
Requirement 2 " Select references to be summarized..." is the object of another
document already submitted.
1 Introduction
1.1 Purpose and content

This paper is a review of the data available that has been used for the classification
of the languages of Mindanao, the Southern, second-largest island of the Philippine
Archipelago. Languages of the Philippines are known to be closely, "genetically"
related and this relationship expands beyond the phonological criteria of the
linguistic Comparative Method, to noticeable resemblances in the lexicon and the
morphosyntax. Still the generally accepted (phonological) subgrouping distinguished
27 distinct microgroups direct descendants of Proto-Malayo-Polynesian in the
"Western Malayo-Polynesian" part of the Austronesian phylogenetic tree, each
originated from proto-languages, - but no intermediate subgroups have been attested
since then by the generally accepted "(phonological) Comparative Method"1, even
though a few scholars have tried hard to demonstrate the past existence of a
reconstructed "Proto-Philippine" language (henceforth 'PPH') different from Proto-
Malayo-Polynesian. In this paper then and unless specified differently, "Philippine
languages" is a geographic notion meaning either "languages natively spoken in the
1
exceptions of interest for us are Central Philippines and Greater Central Philippines (but we counted
them for one, GCP, in the 27), with past existence of related proto-languages attested by the
(phonological) comparative method; and the "Philippines" subgroup, a linkage as attested by a
considerable number of exclusively shared lexical replacement innovations, for which no proof of a
proto-language could be agreed upon to date (but for proto-Malayo-Polynesian of course).

Philippines" (or nearby) or "languages of microgroups with [some or all] members
spoken in the Philippines".
The membership then depends on whether one chose to include all microgroups that
have at least one language spoken in the Philippines, or instead, to retain only the
micro-groups made exclusively of languages spoken in the Philippines. But also, in
theory, on the basis of lexical resemblances one could argue that some micro-group
with no representative language spoken in the Philippine is also part of a Philippine
putative group.
To give a reasonable goal to this review, we steer it towards looking at data

supporting or disproving the arguments given in support of the (non-)existence of a
proto-Philippine language. To make it realistic and feasible within the time and
resources allocated to a term paper for a single class of a graduate curriculum, we
limit it (at least for the detailed examples provided) to languages and micro-groups
represented in Mindanao.
We will start (in Section 1.2) by an introduction to the languages spoken in

Mindanao and their current established classification. Then we will (in Section 1.3)
define "subgrouping" and summarize the most recent debate on the alleged existence
of a proto-Philippine language, to both establish the baseline of facts agreed upon by
all (or most) scholars, and explain the methodological criteria on which the two
camps (pro-PPH and their opponents) disagree - as we will organize our research
along those criteria.
We will then look for available data that could support or at least contribute to
support the following research questions (more explanations will be given in the
"Methodological" Section 1.3), along the following chapters:
- Chapter 2 : phonological innovations - what are, compared to proto-Malayo-
Polynesian (henceforth 'PMP'), the innovations identified in each subgroup - and
are they shared, or exclusive (at least within the scope of the study)?
- Chapter 2.4: lexical innovations - can we find interesting patterns related to
Mindanao, in the sets ("replacement innovations") gathered by Zorc and Blust?
- Chapter 4 : morphological innovations - as there is much less published
comparative data on this and as this was never involved in the "Philippine
resurrection" dispute, we will limit ourselves here to looking at data available on
the verbal inflection systems (limited to voice and aspect) and if we find enough
data, comment on the found differences and resemblances.

1.2 Current classification of the languages of Mindanao
This section is an overview of the languages spoken in Mindanao and their current
established subgrouping.
We refer to Lewis, Simons, & Fennig (2016) (hereafter "Ethnologue") as the "official"
source of current, generally accepted phylogenetic classification of the languages of
the world. Table 1below is a summary of the language micro-groups shown there to
be direct "descendants" of proto-Malayo-Polynesian (we underline), that have at
least one member-language spoken in the Philippines (as per an extract that we did
of the Ethnologue in 2017, so there may have been a few minor changes since then).
The numbers in the table are of languages in each line x column (micro-group x
country) category. There are 176 languages (natively, indigenously) spoken in the
Philippines (out of 1236 Malayo-Polynesian languages). Five micro-groups are
exclusive to the Philippines (Bilic, Central Luzon, Kalamian, North Mangyan and
Northern Luzon). Four others are "shared" with other countries (Bashiic has one
member-language spoken in Taiwan, 'Yami'; Greater Barito has only 7 languages
spoken mostly in the Philippines but also 28 spoken elsewhere; Greater Central
Philippines languages are spoken mostly in the Philippines (90) but 11 languages are
spoken in Borneo or Sulawesi (9 Gorontalo-Mongondow in Indonesia, 1 Danao and 1
Palawanic in Malaysia) ; and Sangiric languages are spoken in Sulawesi (Indonesia)
but for one Philippine language, 'Sangil'. There are also 2 "unclassified" Malayo-
Polynesian languages spoken in the Philippines (but not in Mindanao).
Table 1 Philippines language microgroups
"Philippine" subfamilies Where

Indonesia Madagascar Malaysia Mayotte Philippines Taiwan Total
Bashiic 2 1 3
Bilic 5 5
Central Luzon 10 10
Greater Barito 15 11 1 1 7 35
Greater Central Philippine 9 2 90 101
Kalamian 2 2
North Mangyan 3 3
Northern Luzon 54 54
Sangiric 4 1 5
Unclassified 2 2
Grand Total 28 11 3 1 176 1 220
We also show in Table 2 the subgrouping of the Malayo-Polynesian languages not

spoken in the Philippines. Of course the four microgroups listed above as "shared"
with the Philippines are here again (Bashiic, Greater Barito, Greater Central
Philippines and Sangiric) but the number of languages account only for the ones
spoken "elsewhere".

What is important here is that Table 2 shows 23 Malayo-Polynesian micro-groups (4
spoken partly in the Philippines and 19 exclusively elsewhere); this together with 5
micro-groups exclusive to the Philippines shown in Table 1, make up a total of 28
micro-groups with reconstructed proto-language (plus 4 unclassified languages)
shown to directly descend from Proto-Malayo-Polynesian: no intermediate grouping
has been identified so far on the sole basis of the generally accepted phonological
comparative method, even though tens of senior and respected linguists are or have
been working on this; the intermediate groupings such as "proto-Philippines" that
have been proposed (particularly, by Zorc or Blust) are established on the basis of
lexical arguments and then may be explained as originating from dialect clusters or
"linkages" rather than from a proto-language that could hardly be distinguished
phonologically from proto-Malayo-Polynesian.
If we take aside "Central-Eastern-Malayo-Polynesian" which "mothered" 721
languages (including the Oceanic subgroups), we are left with 27 micro-groups and
4 unclassified languages, grouped "geographically" under the name "Western
Malayo-Polynesian", that is academically not considered either to have originated as
a proto-language (at least none has been attested so far as per the generally agreed
rules of the Comparative Method), but rather as a "Dialect Cluster" or a set of such
clusters, notion now widely accepted that we will come back to in Sections 1.3 and
Chapter 2.4.
Finally, remembering that the scope of our paper here is Mindanao only, we give
details below of the number of languages in Table 1 that are spoken primarily in
Mindanao, and of the micro-groups / proto-languages from which they are currently
considered to originate from. For this we had to split further "Greater Central
Philippines" that spans the three regions (Luzon / Visayas / Mindanao, from North
to South) of the Archipelago, and even further a subgroup of it, "Central
Philippines".
Table 3 shows the number of languages of each subgroup spoken in each region
(Luzon, Visayas, Mindanao) of the Philippines; Greater Central Philippines is spoken
in all three regions, so please Reader wait until the next table to see where with
more detailed subgroupings...
Then when we "open the box", we have a similar situation (in Table 4) with "Central
Philippines", a subgroup of Greater Central Philippines but also spoken in all three
"regions" so we will need to further drill down...

Table 2 Non-Philippine Malayo-Polynesian languages subgrouping
non-Philippine MP languages: 1060

Central-Eastern Malayo-Polynesian 721
Western Malayo-Polynesian (21 families): 339
North Borneo 93
Celebic 64
Malayo-Chamic 55
South Sulawesi 30
Greater Barito 28
Land Dayak 13
Northwest Sumatra-Barrier Islands 12
Greater Central Philippine 11
Javanese 5
Minahasan 5
Sangiric 4
Bali-Sasak-Sumbawa 3
Lampung 3
Madurese 2
Moklen 2
Sundanese 2
Bashiic 1
Chamorro 1
Nasal 1
Palauan 1
Rejang 1
Unclassified 2
Table 3 Regions where Philippine MP languages are spoken...
Philippine Malayo-Polynesian 176 Luzon Visayas Mindanao Subfamilies? (Other comments)

Greater Central Philippine 90 7 subfamilies
Northern Luzon 54 x 2 subfamilies, Arta, Ilokano
Central Luzon 10 x Sambalic (8), Pampangan, Sinauna (1)
Greater Barito 7 1 x Sama-Bajaw (7) - Inabaknon
Bilic 5 x -
North Mangyan 3 x -
Bashiic 2 x Ivatan
Kalamian 2 x - (in Palawan)
Sangiric 1 x Northern (1)
Unclassified 2 x Katabaga, Villa Viciosa Agta
Table 4 Regions where Greater Central Philippine Languages are spoken...
Greater Central Philippine 90 Luzon Visayas Mindanao Subfamilies? (Other comments)

Central Philippine 52
Manobo 15 1 x Kagayanen in Palawan
Palawanic 7 x +1 in Malaysia
Subanon 6 x
South Mangyan 4 x
Danao 3 x +1 in Malaysia
Umiray Dumaget 3 x
Gorontal-Mondondow 0 + 9 in Indonesia / Sulawesi

Finally "opening the box" of "Central Philippines" gives us (almost) unequivocal
regions assignments (Table 5) - almost, because Bisayan languages which are spoken
mostly in the Visayas region, have also 4 offsprings spoken in Mindanao.
Table 5 Regions where Central Philippine languages are spoken
Central Philippine 52 Luzon Visayas Mindanao Subfamilies? (Other comments)

Bisayan 25 x 4 4 "South Bisayan" (Butuanon…)
Bikol 11 x
Mansakan 7 x
(blank) 6 2 4 Ayta (2) / Ata, Binukidnon (2), Sulod
Tagalog 2 x
Mamanwa 1 x
Then we summarize the last three tables into Table 6 - showing finally the first
upper micro-group that each of the languages spoken in Mindanao belong to. In this
table the acronyms GCP for Greater Central Philippines and CP for Central
Philippines remind us of the hierarchy detailed above.
We will now refer to the micro-groups in Table 6 to drive our data quest in the next
sections.
Table 6 Malayo-Polynesian languages microgroups spoken in Mindanao
Philippine Malayo-Polynesian 176 Luzon Visayas Mindanao Subfamilies? (Other comments)

GCP Manobo 15 1 x Kagayanen in Palawan
Greater Barito 7 1 x Sama-Bajaw (7) - Inabaknon
GCP CP Mansakan 7 x
GCP Subanon 6 x
Bilic 5 x -
GCP Danao 3 x +1 in Malaysia
Sangiric 1 x Northern (1)
GCP CP Mamanwa 1 x
GCP CP Bisayan 25 x 4 4 "South Bisayan" (Butuanon…)
In order to do this, we also show in Table 7 the main references from the authors
who are said (according to Hammarström, H., Forkel, R., Haspelmath, M. & Bank, S..
2017, henceforth Glottolog) to have reconstructed the related proto languages.

Table 7 Mindanao micro-groups - main references
Philippine Malayo-Polynesian 176 Luzon Visayas Mindanao Classification

GCP Manobo 15 1 x Elkins 1974; Harmon 1977, 1979
Greater Barito 7 1 x Sama-Bajaw: Pallesen 1977
GCP CP Mansakan 7 x Gallman 1979
GCP Subanon 6 x Lobel 2013
Bilic 5 x Savage 1986
GCP Danao 3 x Allison 1979a
Sangiric 1 x Sneddon 1984
GCP CP Mamanwa 1 x Gallman 1979
GCP CP Bisayan 25 x 4 Gallman 1979, Zorc
GCP Greater Central Philippines 90 x x x Zorc 1986, Blust 1991
GCP CP Central Philippines 52 x x x Zorc 1977
1.3 Methodology: two kinds of subgroups

Robert Blust, a respected scholar who has dedicated his career to studying
Austronesian languages, has initiated in the late 1980's a huge data collection effort
of word lists of the languages of the Philippines and many other Austronesian
languages (e.g. in Blust 1989). His main predecessor in the lexical field is a
pioneering work by Zorc (1986) who had also gathered 121 "Philippine innovations
(replacement etyma)" (23 "widespread", 98 "selective"). Unfortunately, Blust when
promoting his work (and Zorc's) triggered a dispute, arguing that his lexical
evidence was the proof of the past existence of a "proto-Philippine" languages. He
was immediately criticized by non-less prominent colleagues, particularly Reid and
Ross, who reminded him that proto-languages could be attested only by phonological
exclusive innovations, of which none has been recognized so far for the Philippines
alleged subgroup, that would lead to a reconstructed language different from proto-
Malayo-Polynesian. Still both camps agree that in addition to the direct derivation
by divergence, an alternate way to the genesis of a group of languages exist, called
"linkage" (see more on this in 1.3.2 below); surprisingly Blust accepts this, so we
think that the debate is minor and only a terminological one - Blust's lexical work is
outstandingly significant and useful, and proves that Philippine languages have
shared characters distinct from other Western Malayo-Polynesian ones and hence
have a common origin and common early history; but Blust persists to use the term
"proto-language" to designate what most of his colleagues prefer to call "dialect
cluster" or "linkage" (Blust uses both terms...): a group of people speaking diverging
(still close) dialects all derived from proto-Malayo-Polynesian, with overlapping
phonological innovations, that soon further evolved, diverged and crystallized into
different proto-languages that keep those overlapping origins (so that no subgroup
can be identified within the whole set on the basis of exclusive innovations).

This debate is as old as August Schleicher (1821-1868) who first proposed to
develop "family trees" for languages on the basis of the (phonological) comparative
method, and Johannes Schmidt (1843-1901), who later criticized their limitations
and suggested that in some cases, a "wave theory" was more appropriate to account
for the evolution of languages in some given areas, as a result of conflicting
divergence (regular and generalized sound changes) and convergence (borrowings
and irregular sound changes mostly reflected in the lexicons, and guided across
dialect clusters by prestige or sociological factor). But because our paper is focusing
on the Austronesian world, we will first (in subsection 1.3.1) summarize the
application of these concepts in the specified (Malayo-Polynesian) scope.
Because the debate is more about the value of lexical than on the one of
phonological data, and also because lexicostatistics was first used (by Dempwolff
then Dyen), we will also (1.3.2) summarize the generally accepted role of
lexicostatistics and its successors in the historical methods that aim at studying the
relationships between languages, and quote some past results related to the
Philippine grouping.
Finally in subsection 1.3.3we will briefly summarize the arguments of both camps
involved in the "resurrection" debate, and in section 1.3.4 comfort our opinion on
the same with the surprising help... of Blust himself.
1.3.1 AN wave theory: Ross (1994) on "linkages"

To shed some light then on this debate, we will summarize the work of Ross (1994)
that explains that two sorts of grouping can be observed in the Austronesian
Languages Phylogenetic trees, in line with the Schleicher / Schmidt's opposition.
Ross is credited to have detected the past existence of "dialect clusters" or "linkages"
in the Austronesian Phylogenetic tree in 1988 in his "Proto-Oceanic and the
Austronesian languages of Western Melanesia" article2 - but has generalized it to
other nodes of the tree in 1994 in "Some Current Issues in Austronesian Linguistics"
that we summarize here. The same explanation was complemented later in Tryon
[2006], both texts are a solid baseline to be read by anybody who studies the history
of Austronesian languages.
This article introduces a contrast in the ways linguistic-phylogenetic trees are
constructed, with some nodes resulting from dialect differentiation (along Johannes
Schmidt's "waves" theory) and others by separation (allowing proto-language
reconstruction along the original "family tree" concept from August Schleicher).
2
In Ross (1994:45), the author infers that he may not have invented the concept as he states that
"Pawley-Green (1984) use the terms network-breaking and radiation in senses similar to our dialect
differentiation and separation"

Conventionally in the AN tree, and because the general direction of the AN
colonization was south-eastwards (except for Madagascar, Western Indonesia and
the Malay peninsula), the left-hand nodes represent "post-migration left-over" people
and dialect clusters, whereas the right-hand ones represent proto-languages
differentiated by separation following emigration. Pages 48-49 the author further
associates this contrast to the one between using phonological reconstructions (of
proto-languages) and lexical comparisons and detection of "exclusive innovations"
(within a dialect cluster). This is illustrated in detail with comments of the overall
tree, of which we show here below (Figure 1, using a subtree quoted by Blust!) only
the top 5 proto-language levels PAN / PMP / PCEMP / PEMP / P-Oceanic (pp.67-95,
the core of the article; then Ross who was then a specialist of Oceanic languages
provides several additional levels under the latter node).
Figure 1 Austronesian Family Tree by Blust (2009:2)
Particularly and for example, the MP "Proto-Malayo-Polynesian" node has two

dependents - the right-hand one CEMP is also deemed to be a proto-language
(PCEMP, "Proto-Central-Eastern-Malayo-Polynesian")3 and as such it is referenced in
the Ethnologue "official" tree - while the left-hand one WMP, "Western Malayo-
Polynesian", is a convenient way to group the 26-plus micro-groups (please refer to
section 1.2 above) for which proto-languages have been "proven", but could not be
proven to have been itself a single proto-language distinct from PMP; and for this
reason even if shown in the above tree, it does not appear in the Ethnologue tree.
Instead, the WMP "subgroup" is explained by the development of a cluster (or set of
clusters including the Philippine one) of dialects of PMP internally sharing
overlapping innovations before they started to diverge into mutually unintelligible
3
Though according to the author it is the least-well attested proto-language of the tree - incidentally,
Blust is the one who has "invented" it (1993, 2009) - but still it is accepted and represented in
Ethnologue)

languages. "We may conveniently use the term linkage to refer to a group of
languages which have arisen by dialect differentiation." (Ross 1994:46).
The Philippine group then identified successively with lexical support by Dempwolff,
Dyen and many others, and more recently comforted with more examples (of
replacement innovations) by Zorc and Blust, is itself deemed according to these
definitions to have been a "linkage", within the wider WMP set of microgroups "left
over" after the PCEMP migration.
Blust shows in both his original paper (2019) and in his "response to comments"
final paper (2020) that he accepts and understands this concept of "linkage" (he uses
the word several times); he even gives the following example (simplified
here)(2020:452) - assume 4 micro-groups 1 to 4, each defined by 3 phonological
innovations out of 6:
Microgroup 1 - innovations A, B, C
Microgroup 2 - innovations B, C, D
Microgroup 3 - innovations C, D, E
Microgroup 4 - innovations D, E, F
Then it is impossible to find any intermediate group between the proto-language
that would be reconstructed for all four, and any subset of the four. If for instance
you decide that 1 and 2 form a subgroup on the basis of sharing B and C, you
immediately run into a conflict with another linguist who would claim that instead,
2 groups with 3 because they share C and D.
1.3.2 From lexicostatistics to the ACD to... lexicostatistics

Since Johannes Schmidt then - and for the AN family, at least since Ross 1988 and
the study of Oceanic languages (or maybe even earlier, when Blust was working in
1974-1983 on Eastern Austronesian later called CEMP, CMP and EMP languages),
linguists have been aware that the phonological comparative method was not
yielding clear results in the case of "waves" or of "linkages", and aware as well that
the study of lexical changes could bring additional hints where the fundamental
consequences of the law "sound changes are regular" fail to provide evidence. The
study of the lexicon of languages has several interconnections with the comparative
method. First, what we compare in the method are "cognate sets" - across languages,
pairs of form x meaning that are deemed to have either the same form or the same
meaning - the "forms" being actual words ("reflexes" of a proto-word to be
discovered) of the languages being compared. Second, if we were in a first step
collecting cognate sets gathering reflexes of each of the 1,200 Austronesian
languages still alive with the objective to classify them by applying the comparative

method, we would not know where to start: lexicostatistics as a second step would
help us identifying languages that are deemed closer to each others because of the
larger number of cognate sets in which they participate (for the same meaning, the
associated forms are 'obviously' cognates, forms that may be different but derived
from the same "proto-form" along some known universal phonological change; the
same forms with different meanings may not need to be eliminated or split if they
are only traces of a 'semantic shift'). Also scholars found that sometimes identifying
apparent reflexes of some putative cognate was not enough to assess the genealogy
of a given language, because a specific form could have been inherited by borrowing
rather than by sound change from a parent's proto-form; conversely (and this is an
important argument of Blust in his "resurrection" dispute), comparing the lexicons
(inherited or innovated, not borrowed) could help prove the connections within a
dialect cluster or linkage where the phonological comparisons could not help to
define exclusively the characteristics of a proto-language or family.
In short, we will then briefly summarize the characteristics of the lexicostatistics
method, and its most discussed advantages and pitfalls (1.3.2.1); then we will
comment on two extensions deemed to correct the pitfalls, (1.3.2.2) the widening of
the cognate sets base, and (1.3.2.3) the improvement of computations.
1.3.2.1 Lexicostatistics - and its application to the Philippines

Lexicostatistics (according to Greenhill & Gray 2009) has been developed by Morris
Swadesh in the 1950's, whose name is well known by Linguistic students thanks to
the "Swadesh's list" (of 200 every-day-life meanings deemed to not be affected by
borrowings) that is still used sometimes to sketch reconstructions of putative proto-
languages (the trend for the latter task is to use wider lists of at least 300 items), but
according to the eponym article in Wikipedia (unfortunately there is no reference
that I could check for this claim) , it is said that the first application of similar ideas
was by Dumont d'Urville... on Oceanic languages!
Once forms have been collected for the set of 200 meanings for the languages (or a
subset) of a putative family to be analyzed, "distances" are calculated for each pair of
languages as the percentage of meanings that have "cognate" forms - the "cognacy"
judgement needs to be done by linguists who are aware of the possible and plausible
sound changes. On the basis of these distances, groupings and a hierarchy can be
determined, deemed to give indications of an hypothetic phylogenetic tree - that
then needs to be checked rigorously, cognate by cognate, sound by sound, for
confirmation of the regular sound changes that occurred and explain the language
filiations (and if possible prove it by reconstructing a proto-language).

Dyen is the scholar most frequently quoted for the application of the method, and he
worked on both the Indo-European and, coincidently, on the Austronesian families.
For instance, according to Blust (2020:154), "the first explicit effort to justify a
Philippine subgroup was that of Dyen [1965:29-33, 38], who found lexicostatistical
evidence for a 'Philippine Hesion', which includes most, but not all, languages in the
Philippine islands,..."
Lexicostatistics has two main limitations or defects. One of the pitfalls most often
associated to Lexicostatistics is in fact the one of an associated method,
"glottochronology". It has been shown that using an hypothetical fixed "speed" of
change (as glottochronology was doing) yields hypothetical dates for the existence
of proto-languages that do not match the evidence from archeological data on the
settlement of people in the islands of the Pacific. The second limitation is that
lexicostatistics counts the "changes" of form for a given meaning but not the
direction of the change (is a cognate retained - and the other, non-cognate forms for
this meaning are the innovations - or is it the innovation, then the others are the
original, retained form). Those limitations have been corrected by the work of
Greenberg and his associates, described under 1.3.2.3 below.
Another limitation, the (small) size of the samples (word lists) used for the
calculation, is a "virtuous" one rather than a defect - it is known that borrowings can
be a reason of language change, that obviously does not prove genetic relationships.
This is the reason why the Swadesh list is limited to 200 "basic vocabulary" words,
deemed to rarely be borrowed. But then maybe hints are lost that could demonstrate
relationships such as "linkages", for which the 200 words in the Swadesh list may
not bring enough evidence. This is what Zorc then Blust tried to do, and that we
now turn to.
1.3.2.2 Comparative word lists and dictionaries

Glottochronology gives a first idea of the possible groupings of the languages
studied, but does it with a limited number of words. Zorc and Blust have tried to
expand the number of Austronesian words collected. Zorc (1986) started this work
in the early 70's (p.147) at about the same time as Blust (Zorc cites p.150 Blust's
"Austronesian Ethymologies II" - see also Blust 1989, "Ethymologies IV"). This work
is a tedious one as it implies to make a judgement on each "form-meaning" pair
collected, and to classify them - first to eliminate borrowings, second, to assess the
"quality" of each word to contribute more or less to the evidence of a past grouping
(of micro-groups of lower level already proven by the comparative method). Zorc
(1986) describes clearly the classification and criteria that he uses. Let us quote him:

"interpreting each innovation requires isolating the types outlined above:
widespread, selective, contact, or borrowed. Furthermore, the quality of an
innovation must be assessed. I suggest the following measures as a rule of thumb
to weigh the innovations proposed in the various tables (...)." (pp.154-155)
Hence contrary to the comparative method, it is not a scientific method, in that it is
not repeatable: the application of the criteria is still a matter of judgement and also
depends on the number of languages observed (if languages outside of the
Philippines are added to the set, more sets might need to be found that are not
exclusive of the Philippines and reduce the strength of the evidence). Two
researchers might come to different conclusions regarding the inclusion or not of a
given reflex in the list of "shared innovations", even though Zorc's criteria are clearly
described. The review by Blust of Zorc's work, leading to discarding a significant
proportion of Zorc's sets, is itself a clear manifestation of the probability of different
judgements between scholars on the same cognate sets. Still, as Blust argues in his
2019 paper, it is this time (contrary to the Swadesh list) the number of accumulated
observations that makes the argument attractive. As Blust writes (2019:p.163-4),
"the number of proposed Philippine lexical innovations cited in this paper is very
large, but many of them could, in principle, have external cognates." He adds "For
this reason, some scholars are averse to using lexical data for subgrouping. However,
this attitude is unjustified" But he does not produce any demonstration of this, but a
weak statement that "science is inherently probabilistic" (2020:461). Bottom line,
Blust is wrong in his epistemologic argumentation but correct in the likelihood of
the conclusion based on the "statistical judgement"; and the conclusion of both Zorc
and Blust - that a Philippine group has maintained cohesion that unites and
differentiates the languages later issued from it from the other Western Phlippine
subgroups -, confirms results already obtained earlier by the Glottochronology, and
confirmed later again by Greenhill's "bayesian" methods, to which we turn in the
next section.
Zorc (1986) had already identified 23 "widespread" Philippine exclusive innovations
(his Table 1, p.161), and 98 "selective" ones (his Table 2, p.162). But note that in
Table 3 (p.165) he had also identified 79 "proto-hesperonesian innovations",
"linking" the Philippines and Western Austronesian languages. This is
understandable once one accept the explanation of a dialect cluster or "linkage":
some languages (WAN but not PH, spoken by groups that left the Philippines early
in the "fast train" process) differentiated themselves by separation from the linkage
of dialects that started to develop among the proto-Austronesian who colonized the
Philippines, before the bulk of the remaining (then PH) people in the dialect cluster

continued to differentiate themselves, first staying in the cluster and developing
more innovations - this time PH-specific, then starting speaking different languages
proto-forms of the microgroups known today in the Philippines.
Blust (2005, 2019, 2020; Blust & Trussel 2015) have continued the work, their
"Austronesian Comparative Dictionary" (ACD) is still being enriched. In his 2005
paper, Blust writes: "Zorc (1986) [exploring more than 180 PMP languages]
presented 98 proposed lexical innovations in support of Proto-Philippines. External
cognates have since been found for 12 of these (Appendix 1). However, more lexical
evidence supporting Proto-Philippines is now available. Appendix 2 lists 241 lexical,
morphological and semantic innovations not in Zorc (1986) which appears to be
confined to members of the proposed Philippine group."
And in 2019, "the number of likely lexical innovations defining a Philippine
subgroup of AN languages has vastly expanded, and at the time of writing (5/28/19)
stands at more than 1,740. (...) [But] most affixed bases in the ACD are excluded
from the count given here, eliminating nearly 500 potential comparisons." (p.162)
1.3.2.3 Bayesian Phylolinguistics

Expanding the Swadesh list like Zorc and Blust have done conforts the evidence that
at some point of the history of the Malayo-Polynesian languages, the ancestors of
Philippine languages were grouped together (be it as one unique proto-language, or
as a dialect cluster with overlapping sound changes that can still be reconstructed
today); this responds to the criticism of glottochronology on the limited size (in
terms of cognate sets) of the evidence. The other criticism was based on the
limitations of the algorithms, 1. no distinction was made in the data between
retentions and innovations (the direction of the changes was not taken into
consideration - only the fact that for one meaning, two different - or more - cognate
sets were found across the language sampled); and 2. glottochronology was using a
unique constant to modelize the speed of change (e.g. one new language every 700
years). With the increased power of computers, more sophisticated algorithms were
developped to fix both problems - a class of such algorithms is called "Bayesian
Phylolinguistics", and a scholar, Greenhill, has participated in experiments that focus
on the Austronesian family. Greenhill, Currie and Gray (2009), Greenhill and Gray
(2009), and Greenhill, Drummond & Gray (2010) best describe those attempts, that
again confirm the results of Dyen and later of Zorc and Blust - for instance "the
grouping of the Philippine languages is strongly supported by the data (0.99)."
(Greenhill & Gray 2009:10). But also (p.12), "the split between the Northern- and
Meso-Philippine languages is strongly supported (1.00)" This is a claim from Reid

that Blust is mocking in his 2019 and 2020 articles - and even a claim of Zorc
(1986), who actually even distinguishes even "Proto North Philippine" and "Proto
South Philippine" - only on lexical basis though (which is also the basis of
lexicostatistics and Bayesian Phylolinguistics). Note that within the Philippine
group, South Philippine and Gorontalo-Mongondow are also brought up by Greenhill
& Gray (2009) - while instead Sama-Bajaw is shown as an outsider ("only" 0.93) - as
appear in the following snapshot (Figure 2) of the tree developed p.12 of the article.
But Sulawesi is shown to have separated even well before that (and maybe
Minahasan, included in Blust (2019)'s study, not shown here but that could well be
part of the "Sulawesi" cluster).
Figure 2 Bayesian Tree, Philippine portion (Greenhill & Gray 2009:12)
The data converge then on the Philippine grouping subject (with the "marginal"
exclusion of Gorontalo-Mongondow, that Blust placed in GCP); but the items where
the data do not converge are also instructive or worth researching. For instance,
according to Greenhill & al. 2010:
 "The Sangiric language subgroup is placed as a higher-order grouping within
the Western Malayo-Polynesian languages. The Sangiric languages are
located in the Sulawesi region but should be a primary branch of the
Philippines family. Our placement of Sangiric as a deeper group within the
Western Malayo-Polynesian linkage may either reflect contact-induced
change with neighboring Sulawesi languages, or it may reflect the repeated
parallel drift that has occurred in Sangiric and other Sulawesi-area
languages" (p.6).
We come back to this in Section 3.2.2.3 below when we focus later on the
lexically-induced subgroups of Mindanao and the South Philippines.
 "In our data, the major culprit of model misspecification is likely to be
linguistic borrowing - this is probably the cause of at least 21/25
misplacements." (p.4). This confirms the injunctions to detect and eliminate
borrowings from word lists (and due to the difficulties of the exercise, the
fact that lexical analyses are difficult to reproduce and require large number

of instances to compensate with the risk that part of them are incorrect or
not dependable)
 "the phylogenies do not recover two previously proposed subgroups: Western
and Central Malayo-Polynesian. Western-Malayo-Polynesian used to be
considered one of the major subsets of Austronesian, however recent opinion
holds that these languages actually form a number of primary branches
within Malayo-Polynesian. (...) Therefore, rather than being incorrect, the
Bayesian language phylogenies are reflecting this uncertainty in the language
subgrouping". This means then that if WMP was never a dialect cluster, the
subgroups that compose it today are issued from one or several dialect
clusters (including the Philippine one) and possibly of micro-groups that
differentiated separately and early enough to not have been involved in the
development of such intermediate "cluster" subgrouping. This is also we
assume why Ethnologue have decided to not represent this kind of (cluster)
grouping in the phylogenetic trees and to base those on the only subgroups
attested by exclusive phonological changes (hence, proto-languages). In that
sense, "Philippine" is a group with no attested proto-language (as it was
instead a cluster of dialects with overlapping isoglosses).
1.3.3 The "resurrection of proto-Philippine" debate

To promote his outstanding "ACD" enterprise, Blust wrote in 2019 an article
pleading in favor of "the resurrection of proto-Philippine". He chose his title by
reference to an article that Reid had written in 1982, "the demise of proto-
Philippines".
As already explained - or suggested below -, I think both are right. In "The demise of
proto-Philippines", Reid (1982) was suggesting that if a Philippine "group" had ever
existed, it was not as a group of people speaking a unique proto-language noticeably
different from PMP, hence it could not be called proto-Philippines. His paper was a
speculation on which intermediate subgroupings of the micro-groups already then
identified could be found under the level where we now agree that proto-Malayo-
Polynesian was (same level as subgroups of Formosan languages). We know now
(and he recognizes) that his then-hypothesized tree was wrong - raising the level of
Bilic, lowering the one of Amis, splitting the Malayo-Polynesian supergroup (called
Extra-Formosan) into an "Outer Philippines" and (we assume) "Inner Philippines",
called Malayo-Polynesian. His main criterion for the split "Outer / Inner Philippine"
was already the existence of velar nasal clusters: "if the northern languages of the
Philippines do not share in nasal infixation into root words, those languages that do

share it form a subgroup within Austronesian ('inner-Philippines'), and the northern
languages of the Philippines are not part of that subgroup (hence 'outer
Philippines')" (Reid 1982: 211). Reid does not refer to past work on subgrouping on
lexical basis such as Dyen's - he does not have to do it as his usage of the prefix
"proto-" shows that he is only concerned with the (phonological) comparative
method.
In his 2019 paper instead, Blust insists almost exclusively on the lexical innovations.
We have already agreed above that the mass of data he has accumulated points to
the existence of a Philippine grouping at an early stage of the south expansion of the
proto-Malayo-Polynesian people - and that when the existence of a proto-language
cannot be proven by the comparative method, there is another explanation that
linguists have forged over years of experience of classification of the languages of
the world, called "linkages" or "dialect clusters". We have also mentioned that the
method of detecting groupings through "exclusive [to the members of the group],
shared [by most members of the group] replacement innovations" is not a scientific
method in that it is not repeatable or reproducible (as evidenced by the number of
disagreements between scholars on part of the individual, word level claims); still
we agree with Blust that the accumulation of (non-disputed) words that all point to
the same grouping, are strong enough hint that a Philippine group, cluster or linkage
of proto-Austronesian derived dialects must have developed before they diversified
further into languages, "left-over" or migrated south, east and west from the
Philippines.
The rest of the debate then is of scholars who do not really listen to each over - this
being unfortunately triggered by the polemic tone adopted by Blust in his paper,
particularly his violent attack of an obsolete paper from Reid.
What we can build from there is two-fold:
- on the phonological side, confirm whether or not there is any phonological
innovation that could be specific to the Philippines, compared to proto-
Austronesian: how different would be the phonemes of a putative proto-
Philippine language reconstructed by the comparative method, from the
phonemes reconstructed by Austronesian? We will summarize this in Chapter 2 .
- on the lexical side, look into more details at the possible subgroups within the
Philippine group - North, Central and South - and possibly try to find out if some
of what Blust calls "leakages" could be clue of subgroups expanding out of the
boundaries set by Blust for the Philippine group. Blust seems to demise the idea
of subgroups (particularly North and South to his GCP) but it is strongly
suggested by the work of Greenhill and al. described above, and also by

comments by Ross in his "Comment on Blust's..." (2020), as will be reviewed in
Chapter 2.4.
1.3.4 Blust on the importance of phonological evidence

To conclude, and reinforce our position that proto-languages and linkages are two
different kinds of grouping and that the only way to discriminate them is the
phonological comparative method - let us bring in the few following quotes from
Blust, that are enlightening on the methodological issue if not on the specific case of
the Philippines:
Let us note that Blust himself was well aware of these concepts and of the difference
between a "linkage" and a "family" issued from a "proto-language" - let us remember
for example these two quotes (replace "CEMP" by "MP" = 'Malayo-Polynesian',
"Eastern Indonesia" by "Philippines", and "CMP" by "WMP"='Western Malayo-
Polynesian or "Ph" ='Philippine Languages' and the analogy is striking - I propose
the corresponding relevant replacements below in bold italics inserted in the quotes,
the original text is in the current font):
"Among my major conclusions are the following: (1) The evidence for CEMP
[MP] and for some previously unrecognized subgroups within CMP [within
WMP or PH] is considerably stronger than the evidence for CMP [for WMP or
PH] itself. The latter is united (if that is the correct term) by a number of
overlapping innovations that cover many, but not all of the languages in
question. Together with other information, this distribution of non-coincident
innovations suggests that, at an early stage in the AN settlement of Eastern
Indonesia [of the Philippines], the languages now assigned to CMP [to PH]
formed a relatively isolated dialect chain that still shared well over 90 percent of
its basic vocabulary with languages that were not part of this chain. The type of
subgroup that CMP [PH] seems to exemplify thus corresponds rather closely to
what Ross (1988:8) has called a 'linkage.' (...)" (Blust 1993: 244)
And a few pages lower in the same article, under "4.1 Phonological Innovations":
"Perhaps the most striking feature of the phonological history of the CMP [the
WMP and/or PH] languages is the extent to which similar - and in some cases
quite distinctive - changes are found in many, but not all of the languages. This
pattern of innovation suggests that PCMP underwent a short, perhaps
insignificantly short, period of development apart from other contemporary AN
languages before it began to spread from the Moluccas to the Lesser Sundas
[from the Batanes to Northern Sulawesi]. Many of the changes that are now
widespread in these languages took place after this geographical dispersal, and

were the result of diffusion or, in some case, drift. The CMP languages are thus
believed to form a genetic unit of the type that Ross (1988) has called a linkage
(a relatively non-discrete unit produced by differentiation in situ, in contrast
with a family, defined as a product of separation). The reconstruction of the
protolanguage of a linkage presents special problems. As Ross (1988:8) has
noted, 'Whilst a proto language ancestral to a family has a fairly clear status (it is
the parent communalect that existed before separation), the status of a proto
language ancestral to a linkage is potentially ambiguous. Is it the parent
communalect prior to any dialect differentiation [in our case: PMP], or is it the
language at some time after differentiation but before its component
communalects become mutually unintelligible?' [with the additional difficulty in
our case that the boundaries of PH within WMP are almost impossible to define
on the basis of phonological criteria]." (Blust 1993:263)
2 Phonological innovations
In this chapter we look at the phonological innovations that were discussed in the
references that we have reviewed. In Section 2.1 we give a quick literature review to
capture the most important information we have got through (this literature review
will be partly repeated and completed in Chapter 3 to capture the lexical
arguments). In Section 2.2 we look specifically at what we judge to be the three
main phonological innovations that have been discussed as possible evidence for a
proto-Philippine language (rather than a Philippine dialect cluster), and state our
position. In 2.3 we focus on Mindanao to draw a more specific conclusion, or list of
findings or questions.
2.1 Phonological innovations: literature review

 Charles (1974) has proposed phonological innovations shared by Philippine
languages (compared with proto-Austronesian, not with proto-Malayo-
Polynesian) which according to Reid (1982) are four mergers (from *PAN,
nothing is said on the intermediate *PMP level), none being exclusive to the
Philippines. Blust (2019:154) concurs: "Charles (1974:457) proposed
phonological evidence uniting 'a number of languages of the Philippines, North
Celebes, and North Borneo.' However, most innovations that he regarded as
diagnostic for this asserted group define a far wider collection of languages
(merger of PAN *t and *C, or *s and *c), or involve questionable reconstructed
distinctions (e.g. the contrast of *-ey and *-ay, *t and *T, *d and *D, or *z and
*Z)." Though, Blust (2019) adds "the one merger that may carry weight, that of

PAN *d and *z, is easy to miss in this longer list of proposed innovations that are
entirely without subgrouping value." We will look then into the *d/*z merger in
section 2.2.2.
 Conant (1911:72) in his "RGH Law in Philippines languages" writes the
following: "the Philippines Islands form the center of the speech territory in
which the consonant of the RGH series appears as g. Hence it is customary to
classify as belonging to the Philippine group, not only languages of that
archipelago, but such other speech groups as show the g of that series. Among
the non-Philippine languages of this category are the Duzon and Iranun of NW
Borneo, the Singkan Formosan and the Favorlang of Formosa, the Ponosakan and
Mongondow of the North Celebes, and the Chamorro of the Marianas." Note that
Mondondow has been included by Blust in the GCP, but that on the opposite
Blust opposes Chamorro to be part of the Philippine group - as for the Formosan
and NW Borneo languages, I have not seen them yet quoted from this standpoint
I believe in the any other of the papers I have read to date.
 Llamzon (1975) quoted by Reid 2017 has done a top-down search for reflexes of
PAN. His reconstructions are questioned because he seems to be looking for
retentions more than innovation. I observe that his search (of 52 reconstructed
PAN etyma) is based on 9 languages from only two microgroups of the nine
represented in the Philippines (refer to Table 1 above) - Greater Central
Philippines (Tagalog, Cebuano, Hiligaynon, Waray and Bikol) and Northern
Luzon (Ilocano, Ibanag, Ifugao and Kankanay). Reid (2017) acknowledges that
Llamzon and Martin (1976) extended the list to 100 languages. There is not
much usage we can do of this data as it starts from PAN, not from PMP (as the
main question is to find out is whether there was a PPH ever exclusively
different from PMP).
 Paz (1981)'s approach is bottom up - where Blust (1991) discusses together
reflexes r, l, y and g (of a putative *R, be it PAN or PMP?), Paz discusses (pp.31-
32) *d, *g, *l, and *r. We would need to understand the misfit (*d for Paz vs *y
for Blust - but see below Reid 2018, /d/ seems to be specific to Inati, so Paz may
have overlooked /y/ languages and Blust, Inati). But our experiment briefly
discussed in Section 2.2.3 below confirms to us our first impression from our
Lingg365 paper that several if not many reconstructions in Paz are not reliable
as she does not seem to abide by the "majority rule".
 Reid (1982) suggests that nasal clusters / prefixes are found in the Central
Philippines but not in the Northern Philippines (when they do, in Bontok or in
Ilokano, they result of borrowings) hence there may be 2 subgroups. But it is

unclear which one is an innovation. Zorc 2020 seems to say that the cluster is
the innovation, but Blust 2020 instead infers that the cluster already exists in
PAN and PMP, and adds to the confusion when arguing that he has detected
what he calls a "facultative nasal" (2020:458-9).
 Zorc (1986): his introduction quotes Reid (1982)'s "nasal cluster phonological
innovation" to refute it, but does not refute it on phonological arguments -
instead, he initiates the movement later on joined by Blust, to try and "prove" the
past existence of a proto-language solely on the basis of lexical innovations...
 Blust (1991) in his "Greater Central Philippines" hypothesis opposes a table of
20 consonants to the 17 consonants proposed by Charles (this is complementary
to Reid's criticism then: some of the mergers that Charles has found are probably
wrong - or not: are the mergers from PAN/PMP to PPH, or from PPH to the
reflexes in PH current microgroups?). Furthermore Blust (1991) "revisits"
Conant's RGH law (1911) and find four reflexes of *R in the Philippines,
allowing him to group the micro-groups in four classes and also shows that
borrowings detected on the basis of irregularities in these groups seem to be
directional from "high prestige" languages (as an argument to the leveling
brought by the expansion of PGCP). The next section (Section 5 Linguistic
Diversity in the Central Philippines) shows that the density of languages is low in
the Central Philippines (South of Manila Bay to North of Mindanao) but high in
North Luzon and (South?) Mindanao.
In section 6 "the GCP hypothesis" Blust 1991 says that the GCP languages are the
one that merged "proto-Philippines *R and *g into *g (hence **R > *g). Hence it
is crucial to find out the origin of *R - PAN or PMP - at least Charles 1974 did
not talk of PAN merger with *R - and also, as Blust finds 4 groups in the
Philippine languages (r, l and y on top of g), crucial to understand if the 3 others
result of mergers as well - or does *R have other reflexes in those languages?
 Ross (1994) has a comparative table of proto-phonemes of various
reconstructions of PAN and PMP, and a discussion of the method, criticizing the
usage of symbols such as *R (then one does not know which of the reflexes are
innovations...). Cf. Table 2 p.54...Then PAN is detailed pp.55-64. "Issues in AN
subgrouping" explains the left-hand / right-hand distinction (pp.67 to 94, with
all levels detailed...). Follow PMP p.69-72, cf. Table 5 p.71 (compare with Blust's
table 2 on r l y g ?), WMP p.72-81 (note quote p.73 of Wurm-Wilson 1975
'English finder list of PAN reconstructions' "who have reconstructed PPH on
typological grounds")

Note that Ross (pp.74-75) thinks that 7) Sangiric, 9) Gorontalo-Mongondic are
not necessarily related closely to other Philippine languages
CEMP p.81-, and others down the tree... till p.94
 Ross (2005:12-13) "The Batanic languages in relation to the Early history of the
MP subgroup of AN" is quoted by Reid (2018:6): "Blust's putative Proto-
Philippines is doubted by many scholars - see Ross 2005 for a critique...".
 Pawley (2006) "Origin of the Filipinos: Linguistic and Archaeologic Evidence"
quoted by Reid (2018:6) "likewise notes that from the archaeological evidence,
there was no pause during which a homogeneous Proto-Philippines could have
developed"
 Reid 2017: in "revisiting the position of Philippine Languages in the
Austronesian Family", reaffirms the importance of phonological evidence, and
states (p.10) that "Blust 1990's proto-Philippines is no different from the
phonology of PMP". There was a fast spreading migration from Taiwan through
the Philippines first, but soon to Indonesia, Malaysia and the west Pacific area.
 Reid 2018: in "modeling the linguistic situation in the Philippines", again (p.6)
"there is no reported phonological or morphosyntactic evidence that
distinguishes Blust's Proto-Philippines from Proto-Malayo-Polynesian (PMP),
suggesting that the supposedly unique innovations are the result of the spread of
such forms through the country as a result of various trading relationships over
the centuries, and the cross-subgroup dialectal spread common in networked
languages."
This paper has a great discussion of Negrito's part in the assimilation then
development of the PMP language, starting with Inati people (Figure 4 p.10)
which seems to be a first order branch of PMP, the only one with *R>/d/ (Inati
is in Panay). Similarly Arta is a first order branch of PNorthLuzon, the only one
with Ilokano to have *R>/r/. Proto-North-Luzon seem to be limited to r, l, g
reflexes of *R. (Fig.5 p.10). Again we do not see track of /y/, we would need to
check Blust (1991)'s inclusion of /y/.
 Blust 2019: "some [lexical] comparisons in Zorc (1986) are quite strong and are
unlikely to be eliminated by further research. Nonetheless, these may be
considered an inadequate basis for the claim made, given the frequently repeated
statements that a Philippine subgroup is not supported by phonological or
morphosyntactic evidence (Reid 1982, 2018; Ross 2005). (...) With regard to the
former [phonological] we respectfully disagree." (p.156) Follow section 4.1 "the
phonological evidence for proto-Philippines" - "not robust, but that cannot be
ignored". (it follows p.157): "no Philippine language clearly distinguishes PMP *z

(a voiced palatal affricate) and *d in such stable forms as (...) [jfd underlined].
Although it has beeen lost in many languages outside the Philippines, this
contrast is found in many others (...). Given its retention in so many widely
scattered MP languages it is notable that the *d/z contrast is not clearly
preserved anywhere in the Philippines." (p.157)
But then Blust admits that the contrast has been lost also "in many languages in
the Philippines", why does he cite it as a "phonological evidence for proto-
Philippines"???
Note that in Paz (1981) in her 'tentative correspondence chart' (Table 2) we find
/z/ reflexes only in Ibg, line 8, under column 4.
A possible counter-argument is that Kapampangan have merged *d and *z as
well but still keep 4 nasals, which is contrary to Ferguson (1966)'s rule - Blust
says it instead concurs (but it is not true if it is only Kapampangan...!!).
 Zorc 2020's "reaction to Blust's..." brings the following elements to the
phonological debate:
o "contrastive word and morphemic accent in the Philippines (section 3
p.3) may well be an innovation. Furthermore, the majority of PH
languages support the hypothesis that words of the shape CVC.CV(C)
were accented on the ultima (in contrast to the penult accent of PAN and
PMP)."(p.2)
o "the reconstruction and value of initial PPH *y- has long be ignored"
(section 9 p.22)"
 Reid 2020's response on the phonological argument: there is a long discussion
about Blust's dropping of initial glottal stops in his ACD but we do not find it to
have any value in the PPH dispute. Reid comes back on the nasal clusters that he
claims are separating Northern Philippines from the rest but the explanations (of
both parties) are confused, and it is not either relevant to PPH but instead of the
existence or not of a lower level (split of PH between North and South).
 Ross 2020's "comment on Blust's resurrection..." does not comment on the
phonological theme but merely notes that "Blust proposes that PPH is now (jfd
emphasizing) supported by a single phonological innovation, the merger of PMP
*z and *d."
 Liao 2020's "a reply to Blust's resurrection..." refutes Blust's claim that the *d
and *z merger could be phonological evidence for PPH, in these terms (p.433):
"in addition to the abovementioned potential counterexamples to the claim that
all Philippine languages merged PAN/PMP *d and *z, the validity of this merger
as subgrouping evidence has been questioned not only by Reid (1982) and Smith

(2017) but also by Blust himself before. In discussing evidence for subgrouping
Formosan languages, the Austronesian languages spoken in Taiwan proper, Blust
(1999:43) states 'Extremely widespread mergers, as *d/z, or *N/ñ are ignored as
non-diagnostic..." Based on Blust's (1999:44) study, the following Formosan
languages exhibit the merger of PAN *d and *z: (...) These languages represent
five first-order subgroups (among the ten first-order subgroups) of the AN
language family. Since the merger of PAN *d and *z is considered "non-
diagnostic" for classifying Formosan languages, it should be considered non-
diagnostic for subgrouping Philippine languages, too."
2.2 Phonological innovations across and within PH

Those innovations may be related to two different standpoints, a) claims for PPH
overall ones (departing from PMP) - and claims for (unattested to date) groups (of
attested microgroups) within the Philippine languages. We cannot analyze all those
claims (particularly when they are not consensual) within the reduced scope of this
paper, so we will focus only on the three ones that we judge the most widely
discussed: the g reflex of the RLG Law (Conant 2011), the merger of *d and *z (Blust
1991), and the "prenasalized clusters (Reid 1982). We let aside the "morphemic
accent" stress because it is quoted only by Zorc (2020) in our review, with no data
provided (except maybe by Paz 1981, but with no data of non-Philippine languages
that would help gauging its exclusivity degree)..
2.2.1 The RGH Law

At the time when Conant was writing this paper, the structure of the whole
Austronesian family had not been perceived by scholars. Conant calls the superset he
is looking at "INDonesian" languages. He looks at 36 languages, of which only 2
Formosan (one extinct) and 2 CEMP; all the others are spoken in the Philippines,
Malaysia and Indonesia. The observation of two of the provided tables, reproduced
below as Figure 3, show that
1. the reflex g is not exclusive of the Philippines (Chamorro and Formosan
languages have it too), and in the Philippines, seems rather characteristic of
GCP, with the exception of Ibanag (North Luzon) that has been said to be more
conservative... Indeed the GCP microgroup is recognized to have originated from
a proto-language (rather than from a cluster).
2. in the Philippines, some subgroups have replaced *RGH by a non-g reflex:
 most Northern Luzon languages (as well as Kalamian) have h (only Iloko has r
and Ibanag has g). On the basis of this innovation they would group with

Sangiric (still at the edge of Philippines x Sulawesi), but also Dayak (still
plausible "frontier" in North Borneo) and... CEMP / Oceanic! (Bulu is spoken in
West New Britain - this may be an independent, parallel innovation). Having h
reflex both in the North and the South of the Philippines is not a surprise if they
originate from a dialect-cluster, as Blust has demonstrated that proto-GCP has
later on crystalized and developed as a proto-language (probably issued from the
same cluster) and generated an "extinction" of formerly more diverse languages
in the "(wide) Center" of the Archipelago
 Bashiic and Central Luzon languages have y and would group with Lampung
(Southern Sumatra) and North-West Sumatra (Gayo is spoken in Central Aceh).
We have read in a few articles that the Batanes might have "recolonized" from
the South after the natural first Taiwan-to-Philippines wave of immigration: this
might have been done by people established in Luzon, with another part
migrating to Sumatra and the rest displaced to Central Luzon by the PGCP-
speaking people.
 Bilic has r (and maybe g, if Conant's "Bagobo" is Giangan rather than Tagabawa)
like Ilocano;. ff this were an innovation it would group then (except for parallel,
independent innovations) with languages in NW Sumatra, Malayo-Chamic and
South Sulawesi.
The conjunction of the dispersal of the r, h and y dialects would tend to reinforce
the idea of an original dialect cluster, with dialects having one or the other of the
four possible reflexes already, that spread rapidly, not only to the Philippines but
much further (to North-West Sumatra South Sulawesi, to Guam for Chamorro, even
maybe to West New Britain of Papua New Guinea for Bulu), before crystalizing into
proto-languages, with PGCP (language of one of the g dialect-people) "replacing"
many languages in the "Great Center" of the Philippines and reducing the diversity
there as explained by Blust.
In addition, the presence of 2 Formosan languages in the superset, knowing now
that this was the cradle of the Austronesian people, makes us suspect that g might be
a retention rather than an innovation - but this is only a speculation, we will try and
find out in Ross (1994) how the proto-phonemes (PAN, PMP) corresponding to
**RGH is now represented. In any case this would disprove the hypothesis of a PPH
based at least on this innovation, but of course would remain compatible with the
"dialect-cluster" explanation.

Figure 3 RGH Law: reflexes by source according to Conant (1911)4
Note also that the last part of Conant (1911:82ff)'s paper discusses the "three-fold
origin of the Philippine g: the g's of the Phil. languages may be divided into three
classes according to their origin, namely original g, the g of the RGH series, and that
of the RLD series. Indeed Paz (1981) in her Table 2 "Tentative Correspondence
Chart" shows 3 morpheme correspondence sets were several g are frequent (column
7 must be "the original g", column 8 is seems to be the RGH one, and column 17 the
RLD one. This may be a useful reference to find more reflexes for focalized
complementary studies.
Finally, let us remember that Blust (1991) has a section "the RGH law (...) revisited"
(p.89). He does not question Conant's claims but instead complete the data with
more Philippine languages, and also shows that PGCP was a "prestige" language that
induced a lot of borrowings in the North and the South of the Philippines (hence g
reflexes of PMP *R are likely to be the signal that the etyma where they are found
are borrowings, not regular cognates).
We did not check Zorc 2020 or Blust 2019 hints on the 'long ignored' /y/ but we see
in the tables above that this reflex, shown to be shared within the Philippines by
Central Luzon and Bashiic, is not exclusive of the Philippines as it also appears in
Lampung (South East Sumatra) and North-West Sumatra.
2.2.2 The merger of *d and *z

Blust (2019:154,157) says that this merger (into d? or z?) is part of Charles
(1974:457)'s findings, whose he rejects the other claims. Here then is a summary of
the "*d & *z merger" pros- (2.2.2.1) and cons- (2.2.2.2) arguments.
4
In each pair of columns, the left one is the name of the languages exemplified by Conant - sometimes
I changed the orthography to the modern one; the right one is the highest subgroup under the PMP
family (but for the 2 Formosan ones) in which those languages are classified today in Ethnologue

2.2.2.1 "Pro" arguments
The merger of *d and *z as exclusive to the Philippines is said to have been brought
in by Charles (1974:457), also quoted by Reid 1982, 2017, 2018 who rejects it, and
supported by Blust (2019:154,157):
"the one merger that may [jfd emphasizes] carry weight, that of PAN *d and *z, is
easy to miss in this [Charles'] longer list of proposed innovations that are entirely
without subgrouping value" (p.154)
"no Philippine language clearly distinguishes PMP *z (a voiced palatal affricate) and
*d in such stable forms as *zalan 'path' vs. *duha 'two', or *quzan 'rain' versus
*udang 'lobster'. Although it has been lost in many languages outside the Philippines
[jfd underlines]... (...) Given its retention in so many widely scattered MP languages
it is notable that the *d/z contrast is not clearly preserved anywhere in the
Philippines." (p.157).
Blust adds considerations on the reduction of the number of nasals in a language
deemed to be parallel to the one of obstruents, but we consider that this goes
nowhere as the author has already disqualified himself the *d / *z merger as being
exclusive of the Philippine group of languages that he wants to prove originated
from a proto-language rather than from a dialect cluster: he wrote that the contrast
has also "been lost in many languages outside the Philippines"...
2.2.2.2 "Con" arguments

Liao (2020) is the one who most clearly refutes this tentative by Blust to pretend
that he has identified an exclusive innovation in this *d/*z merger.
First, she points that the merger does not apply to all languages of the Philippines -
for instance, quoting Blust himself: "Blust (1999:36) states that '... there appears to
be some fragmentary evidence that PAN *z is retained as a palatal in Inati of Panay
Island'... If Inati truly shows [this] evidence, the exclusion of Inati will mean the
removal of an important piece of evidence for the diachronic development of PAN
phonemes and/or other properties..." (Liao 2020:428). Also, "the merger of
PAN/PMP *d and *z does not seem to be perfect in Philippine languages - for
instance Blust (p.154, footnote 4) notes that 'an apparent exception is Ayta Abellen
odan 'rain' (< *quzan)..." (Liao 2020:432).
Second and more importantly, she shows that the *d/*z merger, even if it were to
delineate a (reduced) Philippine subgroup, is not exclusive hence not a diagnosis for
a proto-language: "In addition to the above mentioned potential counterexamples to
the claims that all Philippine languages merged PAN/PMP *d and *z, the validity of
this merger as subgrouping evidence has been questioned not only by Reid (1982)

and Smith (2017b) but also by Blust himself before. In discussing evidence for
subgrouping Formosan languages (...), Blust (1999:43) states 'Extremely widespread
mergers, as *d/z, or µN/ñ are ignored as non-diagnostic...' Based on Blust's
(1999:44) study, the following Formosan languages exhibit the merger of PAN *d
and *z: Atayalic,...". Since the merger of PAN *d and *z is considered as non-
diagnoctic for classifying Formosan languages, it should be considered non-
diagnostic for subgrouping Philippine languages, too." (Liao 2020:432-433)
2.2.2.3 Discussion and conclusion on the *d/*z merger

As noted by Liao, and as Blust recognizes himself, the merger of *d and *z even if it
seems widely spread across almost all Philippine languages is non-exclusive of the
Philippine languages, hence should be discarded as a proof of the existence of a PPH
language.
Note that Zorc (1987) has published an article titled 'Austronesian apicals (*dDzZ)
and the Philippine non-evidence" where many examples are given of Philipinne
etyma that should not be used in the reconstruction on the basis of d or z reflexes of
the corresponding PAN / PMA proto-phonemes because of a full set of reasons
(including borrowings) that he illustrates in details.
Note also that in Ross (1994), Table 3 (p.56) shows only 3 Formosan languages (or
microgroups) with a z reflex (or lowe-level proto-phoneme): proto-Puyuma
(reflecting PAN *d2 or *d3, in alternance with retroflex d), proto-Paiwan (for *d2),
and Saisiyat (for PAN *j). Also in Ross (1994), Table 2, p.54, four reconstructions of
PAN have a **Z reflex in addition to the D/d1-2-3, for which Dempwolff's
reconstruction (of PMP) shows no reconstructed proto-phoneme. Finally Table 5,
p.71, shows that indeed both PAN **Z and PMP *Z give Tagalog d (or intervocalic l)
like PAN **d1/2/3 (with intervocalic r); still there is a hint that some other PMP
languages had merged **d1/2/3 to *(z), merger but in the other direction.
Indeed as both Ross (1994) and Zorc (1987) suggest, the reconstructions of PAN and
PMP are still work in progress or still disputed in the details... This makes comparing
a putative PHP with PMP not easy, but still I think the arguments above on the
*d/*z merger are still solid enough to discard it as evidence of an exclusive
innovation that would be an evidence for a past PPH language different from PMP
(but for local dialectal variations).
Finally note that Blust seems to often change opinions... in his "Greater Central
Philippines Hypothesis" (1991:87-88), he opposes his own table of 20 "PPH
consonants" to Charles (1974)'s 17 ones, and the former does include both *d and *z
as separate "proto"-phonemes...

2.2.3 Nasal clusters
Reid (1982) detects a split in languages of the Philippines, between those that have
and those that don't have an inherited prenasalized cluster: most of the Philippines
(?) for the former, and Bontok, Ilokano for the latter (and maybe all the "Northern
Cordilleran" languages, that we assume in Ethnologue to be "Northern Luzon /
Ilocano" and "Northern Luzon / Meso-Cordilleran") - with the justification that the
(rare) reflexes that still have one are borrowings and are not inherited. Of course the
speculative conclusions that Reid draws from there in terms of the format of the
phylogenetic tree (extra-Philippines, raising of Bilic, lowering of Amis...) are
adventurous and even himself has pulled them (officially in his 2020 response to
Blust...). But the possible coexistence within the Philippines of two separate groups
(that may span outside the Philippines) is worth looking at - in relation to the debate
of why PGCP stands in-between languages in the North and South that have some
(lexical) similarities. In his conclusion Reid hints that maybe Bilic languages of the
South of the Philippines, Blaan and Tboli, share the same characteristic as the North
Cordilleran (no -N- nasal infixations) but furthermore, that the other Philippine
languages (those that have -N-) share this "innovation" with Indonesian and Oceanic
languages.
Let us note for fairness that Blust 2020:458-459 wrote 2 pages to refute Reid's
claims, but we found the argument a bit subtle and we have to differ its analysis to
some later research.
2.2.3.1 Experiment 1: Paz and the nasal clusters

We tried to check the variety of reflexes of nasal clusters putatively found in PPH in
Paz (1981) but the experiment did not bring a strong conclusion. We found 39
proto-morphemes in Paz's Appendix 1 with a nasal cluster: 28, 29, 30, 60, 61, 63,
95, 125, 143, 152, 170, 171, 172, 173, 195, 198, 201, 210, 214, 215, 227, 228, 239,
249, 251, 254, 285, 299, 300, 302, 311, 312, 313, 314, 315, 340, 351, 374, 389.
Unfortunately, first, we found that at least 8 reconstructions seem incorrect (29,
201, 239, 302, 311, 340, 389) - the author does not seem to apply the "majority
rule" of the reconstruction methodology, in those cases there are only one or two
reflexes with a nasal cluster - why cannot the author imagine that nasalization
instead is the innovation in these cases, is not it a natural phonological change
process from her standpoint?
But more importantly, even though it seems to be confirmed that the reflexes that
lose the nasal of the "proto-clusters" are mostly in Northern Luzon as hinted by
Reid, the magnitude of the proportions is not outstanding: 7 Ilokano, 6 Itawit, 5

Itbayat (actually this is Ibatan / Bashiic), 4 Pangasinan, 4 Ibanag, 4 Bontok, 2
Kalinga. But there are also 15 languages that have one or two reflexes that lost the
nasal. Of course again this is not significant either and does not disprove Reid's
claim.
2.2.3.2 "Con" arguments

Zorc (1986) and Zorc (2020) contest this idea (Blust 2019 does as well but in so
aggressive terms that we better disregard his comments, and anyway Blust's position
is an extension of Zorc's).
But actually if one reads Zorc (1986) carefully, he does not disprove specifically
Reid's phonological evidence (hence he even seems to confirm it: except for North
Cordilleran and Bilic languages, the other languages of the Philippines share a
characteristic, insertion of a -N- nasal infix in root words, "with Malay/Indonesian
and Oceanic languages". Instead, he starts then the core part of this paper, which is
to start accumulating putative "shared lexical innovations" limited to the Philippines,
in order to claim for the past existence of a specific group of Philippine proto-
languages.
Blust 2020 brings also "con" arguments (to counter Reid 2020), though they are a bit
tricky and confusing - he talks (pp.458-459) of "facultative nasals".
2.2.3.3 Discussion and Conclusion of the -N- criterion

I feel that Reid's findings states clearly that if the -N- infixation is an innovation, it
cannot be accounted as evidence of a proto-Philippine language because it is not
exclusive, it is explicitly said to be shared with Malay/Indonesian and Oceanic
languages. In addition it is not an innovation as it was found by a student of Reid
looking for reflexes of PMP words (if PMP had this infix, then the languages that still
have it in Philippines, Indonesia and Oceanic are retentions, not innovations). In the
latter case, the innovation then would be the loss of the -N- infix - but then as it
seems limited (in the Philippines) to North Cordilleran (and Bilic?) languages, it
cannot be either accounted for an exclusive innovation that would demonstrate the
existence of a proto-language for the whole of the Philippines...
Let us also note that Zorc (2020:9) does not try to discredit further Reid's
phonological findings. Instead, he confirms it: "Tendency towards CVCVC word
structure [without NC] in North Philippines | vide Reid (1982). The introduction of
CVNCVC word structure (nasal incrementation) probably started in the Southern
Philippines, which later became even more prominent in the Indonesian and
Oceanic branches."

2.2.4 Conclusion on innovations from PMP in the Philippines
None of the three innovations discussed could be shown to be exclusive of the
Philippines, hence none can be used to demonstrate the existence of a putative
proto-Philippine language... On the other hand the lexical data gathered by Zorc
then Blust point to the consistency and common history of the microgroups
authenticated as being issued from proto-languages (see the Ethnologue official
classification), but this is academically explained by the concept of "dialect cluster"
or "linkage" abundantly described for instance in Ross (1994) or Tryon (2006) and
many others (even Blust when he was studying CEMP and CMP). The only way
forward to confirm the lack of exclusively shared innovations would be to compare
reconstructions of PMP and of the putative PPH, but it seems according to Ross
(1994) that the reconstructions of PAN and PMP at that time was not yet fully
agreed across the board, so this will probably take many more years of systematic
research and data gathering and verification... For now we consider the "dispute"
started by Blust (2019)'s "resurrection" paper as a terminological one - Blust persists
to call the Philippine (lexically) attested group a proto-language descendant,
whereas most if not all scholars tell him that this is instead exactly the prototype of
what has been called a linkage or dialect cluster by the linguistic community.
2.3 Focus on Mindanao

Could some of the phonological arguments discussed above be used to extend some
of the current generally accepted views on Mindanao languages subgrouping (by
identifying higher level subgroups depending of PMP, if not the Philippine group) -
within the Philippines, or more widely across the Western Malayo Polynesian (if
issued from a dialect cluster, some groups might overlap)?
We remember from section 1 (Table 3 to Table 6) that 4 microgroups direct
descendants of PMP/(putative) PPH have languages spoken in Mindanao - GCP,
Greater Barito, Bilic and Sangiric.
 Bilic is spoken exclusively in Mindanao (and in the Philippines).
As for Bilic, let us note that Zorc (2019) in a paper titled "Klata / Giangan: a new
Southern Philippine Subgroup" claims that Giangan, to date considered a Bilic
language, is actually an isolate, member of Zorc's claimed "Southern Philippine
Subgroup". The paper again is based only on lexical evidence, and quotes not
only Philippines, but also Western Malayo-Polynesian as subgroups of Malayo-
Polynesian, without any mention to the proto- / cluster ongoing discussions. "all
the above [jfd: lexical] information appears to validate Klata, Bagobo-Klata or
Giangan as an independent language, a single node on the Southern Philippine

family tree, coordinate with Bilic, Tboli and Teduray and ultimately the GCP
languages." (op.cit., p.48).
Savage (1986) has reconstructed proto-Bilic - refer to section 4.5.1below.
 Of the 35 Greater Barito languages, 7 Sama-Bajaw are spoken in the Philippines -
5 are from the "Sulu-Borneo" descendant (this micro-group has also one
representative in Indonesian and one in Malaysia), 1, Yakan (isolate descendant)
is spoken in Basiland Island, and 1, Inabaknon, is spoken in Northern Samar, not
in Mindanao. The literature refers to "Sama-Bajaw" as a nomad people.
 Sangiric has five dependent languages, 1 North-Sangiric 'Sangil' spoken in the
Philippines, and the 4 others in Indonesia / North Sulawesi (2 "North" 'Sangir'
and 'Talaud', and 2 "South-Sangiric". Sneddon (1984) has reconstructed proto-
Sangiric (refer to section 4.5.2 below).
 GCP is spoken mostly in the Philippines (90 of 101 languages), and has four
direct-issued microgroups spoken in Mindanao - Subanon, Manobo, Danao (all 3
mostly spoken there, only Manobo has one of 15 members spoken in Palawan) -
and CP mostly spoken in South Luzon and the Visayas, but in turn with 2 direct-
issued microgroups spoken exclusively in Mindanao - Mansakan, Mamanwa.
From our standpoint (phonological so far), as we are looking for clues of the
existence or not of a proto-Philippine language, GCP has no value: GCP is
accepted to be a micro-group ("registered" by the Ethnologue) and is "contained"
in the scope claimed to have been the one of the Philippine group (including the
Gorontalo-Mongondow micro-group of Indonesia), so cannot disprove the
hypothesis that the Philippine subgroups might have been issued from one
(higher-level) proto-language.
So in this section we should look primarily at languages of Mindanao that are not
part of GCP, to uncover any clue of potential grouping with other languages (not
spoken in Mindanao or even in the Philippines) - namely then, languages of the
Bilic, Sangiric and Greater Barito microgroups. And we will do so along the
phonological specificities already discussed above for the Philippines languages,
specificities that are our overall scope in this paper.
We follow the same plan as above and review again each of the three putative
phonological criteria:
2.3.1 Reflexes of *R (the RGH Law)

We discussed already in section 2.2.1 "The RGH Law" that according to Conant
(1911), Sangiric (Mindanao, Sulawesi) groups with North Luzon as languages

showing the reflex "h" while Bilic (Mindanao) groups with Ilocano as languages with
the reflex "r".
As for Sama-Bajaw on the other hand, there seems to exist a consensus that they
should not be considered as part of the Philippine subgroup (be it a linkage or issued
from a proto-language). And they are not sampled in Conant (1911)'s list of 36
languages. We do not have data to show what are their reflexes for *R so let this
question open for later research.
2.3.2 The *d/z merger

Sneddon (1984) shows no z in the Sangiric consonants, only d (but also dȝ - could
this be the "lost" reflex for *z?) - hence on this basis Sangiric languages would indeed
belong to the Philippine putative microgroups (we would still have to demonstrate
that this merger is an exclusive innovation, not found outside the putative group).
Sama-Bajau does not seem to have z either (but actually is not considered by tenants
of proto-Philippine as a Philippine languages; this would be a contradiction, showing
that this "d/z merger innovation" then is not exclusive of the Philippine subgroup...).
Our source is Miller (2007) but this source is restricted to "West Coast Bajau", and
also the consonant table shows 4 voiced stops (matching also 4 nasals), b/d/g/dȝ.
We would need to check (in Pallesen 1985? we could not procure it yet) that dȝ is
not a reflex of *z in proto-Sama-Bajaw.
Bilic has only b/d/g (but also only 4 nasals - referring to Blust (2019, 2020)
argument that the number of voiced stops should match the number of nasals. From
this view, Bilic would be a Philippine language then (again, this would not be
enough to demonstrate the past existence of a PPH language, another condition
being that this innovation should have been exclusive - specific of the putative
group).
2.3.3 The -N- prenasalized medial clusters

It is more complicated to assess whether Sangiric, Sama-Bajaw and/or Bilic have (or
don't have) -N- prenasalized medial clusters, because this requires to look into
wordlists, rather than merely to inventories of consonants. [JFD: if I had more time I
would browse Sangiric by Sneddon 1984's and Bilic by Savage 1986's wordlists as I
have them...]. We will come back to this then in "lexical" chapter 3 below. Some
data of reference are:
 table 2 in Reid (1982:205), that has a list of 22 "PPH-reconstructed" proto-words
with such clusters that don't have a nasal in Bontok: we could check them
"against" their equivalent in other PMP reconstructions to see if they were

innovations or retentions in Bontok, and then check their reflexes in some
Indonesian and Oceanic languages to find out if the presence, or absence of such
cluster is exclusive of the Philippines;
 table 2 "proto-South-Philippines" in Zorc (2020:12-15), out of 124 cognate sets,
has 13 with medial clusters - one can try to find the corresponding reflexes at
least in South Mindanao languages; table 1 "proto North Philippines", indeed, of
82 items has almost none reconstructed with a medial nasal cluster (the only
exceptions are *bungtut resulting of the los of a medial schwa, *denden and
*dengdeng that can be explained by reduplication, and *taluntun the only
unexplained one).
2.4 Conclusion
In this section we have reviewed 3 areas of phonological changes that have often be
quoted as being maybe specific of the Philippines but it seems that the evidence is
not really there, there are many exceptions that defy the expected rigor of the
Comparative Method. Even when looking at Mindanao only, it seems that the said
criteria cannot even be used to prove the unity of Mindanao.
3 Lexical innovations
3.1 Literature review

Here we focus on criticisms of Zorc's and Blust's lexical work by other scholars, and
Zorc and Blust's defense.
 Reid (2020) contrary to Ross does not look into the details of Blust's findings but
has two overall methodological arguments.
First he states that the lexical facts described by Blust are best explained by a
dialect network or linkage, in addition to not being evidence of a proto-language
- as seen in the following quotes:
"while tree diagrams and chained language diagrams can effectively model
certain facts about the historical development of languages, they do not
model the kind of extensive networking resulting from trade. In the
Philippines, this has possibly resulted in the kind of data that Blust [1999]
uses to construct his 'Proto-Philippines', large numbers of words that have so
far not been found to have cognates outside the Philippines. Many of these
are probably words that have moved because of trade and been widely
adopted." (Reid 2018:11)

"there are sets of innovations that are restricted in their distribution. Blust
claims that these are due to a dialect chain (...). PMP spread through the
Philippines rapidly, according to archaeological evidence, which Blust
endorses, but claims that the dialect chain was his PPH. 'It is hard to avoid
the conclusion that Northern Luzon and GCP formed the central axis of an
earlier dialect chain stretching the length of the Philippine archipelago and
perhaps beyond' (Blust 2019, pp.219-20). He claims, 'this community formed
a dialect chain from north to south (the natural direction of population
movement in an archipelago the shape of the Philippines)' (p.184). This
implies a PMP dialect chain (whether or not the Batanes Islands were
involved), rather than Blust's PPH."(Reid 2020:386).
Second, he reminds us that Blust's list constitutes "negative evidence" - providing
etyma whose reflexes are deemed to be spoken only in the Philippines cannot be
strictly proven until all non-Philippine Austronesian languages have been
checked... Indeed Blust insists that considered the magnitude of his work, even if
some etyma are discarded there should remain a sufficient number to continue
to strongly point to the past existence of a linkage or dialect network (which
nobody disputes, in fact). Actually Blust himself is aware of this relative
weakness, as his following quote by Reid shows:
"while Blust has provided many sets of cognates (...), he assumes that
because of their quantity they are found only in the Philippines, although he
is careful to state (p.163), 'but many of them could, in principle, have
external cognates.' (...) as Blust (p.164) notes also, 'lexicographic resources
for Philippine languages are far richer than those for many other areas, in
particular Borneo, eastern Indonesia, and much of Melanesia, and this
difference in documentation statistically favors comparisons that are robustly
exemplified in the Philippines but are, in at least some cases, very rare
elsewhere.' " (Reid 2020: 386).
 Ross (2020) in his 'Comments...' provides an in-depth quantitative analysis of
Blust's data and concludes (p.367) that
"the distribution of the innovations cited in support of PPH does not attest
that the Philippine languages form a proper subgroup. A proper subgroup is
ideally defined by one or more innovations shared by ALL its members and
exclusive to the languages of that group. (...) This defines an ideal, and one
that is rarely achieved for large language groups. (...) Nonetheless, one
would expect a few innovations that come close to defining PPH by being
reflected in most of its 17 microgroups. (...) Among the 37 strong

(replacement) innovations only 4 have any claim to be Philippine-wide, and
even these skip some micro-groups." (p.367)
"Instead of a collection of innovations defining PPH, the pattern that the data
present is of smaller groups, intersecting each other in varying degrees [jfd:
hence forming a 'linkage'!...]. From the strong innovations, I calculated the
cohesiveness of all 136 possible pairings of Blust's Philippines microgroups.
(...) The most cohesive pairing to emerge is {2b, 4a} at 64%, that is, North
Cordilleran and Central Philippines: of the innovations affecting 2b and 4a,
whether they occur in both micro-groups or in one of them and one or more
microgroups other than 2b and 4a, 64% are common to both 2b and 4a. This
unexpected pairing is followed by the pairings shown here:" (Ross 2020: 368)
(... follow list of 9 more pairings all with a cohesiveness higher than 40%...
let us note as our focus in this paper is Mindanao the pairs 2. Manobo x
Danao at 60% and 7. Central Philippines x Manobo at 44%. There are even
prominent pairs involving Danao and some Northern micro-groups but rather
than interesting for our focus on Mindanao, are more indicative of the fact
that these innovations were shared within a network located in a smaller part
of the Philippines and that later on expanded over larger territories, creating
"gaps" that given the number of etyma cannot result of independent
innovations." (p.367-368)
 Note in addition that on the lexical side, Zorc (2020) goes even further than
Blust as he now claims the existence of two proto-languages, PNP / PSP, proto-
North- and -South-Philippines... (list of cognates in Tables 1 and 2 respectively,
pp.10-15). This is confusing (GCP is considered part of the "South Philippines" as
attested by the presence of a number of Tgl and Ceb reflexes in Table 2), as until
now most scholars involved in this discussion had accepted that GCP had "sat" in
the middle of the Philippine group dialects, explaining some resemblances
between presently north-north and south-south languages...
3.2 Experiment 2 - more cohesive pairings

We are not armed to take a critical look at Zorc' nor Blust's data, they are experts
and we cannot have at our level any value added nor comments on the data they
gathered. Still, Ross (2020)'s analysis gives us a lead to try a small experiment and
prolong his insight. Ross computed the "cohesiveness" of pairs of microgroups and
found that the most "cohesive" pair in Blust (2020)'s Appendix 1 1222 "PPH
innovations" was 2b and 4a, "North Cordillera x Central Philippine": "that is, of the
innovations affecting 2b and 4a, whether they occur in both microgroups or in one

of them and one or more microgroups other than 2b and 4a, 64% are common to 2b
and 4a".
We observe that Bilic and Sangiric are all missing from the top 10 most cohesive
pairs (while Manobo, Central Philippine, and Danao, other representants of
Mindanao, are there). In addition, Blust's data does not include any representative of
Sama-Bajaw that he says is not part of "his" Philippine group; but instead,
Minahasan, a microgroup of Sulawesi, is included.
We tried then as an experiment to calculate the cohesiveness of pairs involving Bilic,
Minahasan or Sulawesi, to see which of the other Philippine microgroups are the
closest to them.
3.2.1 Revision of Ross (2020)'s calculations

But first when we re-computed the cohesiveness of the pairs manifested in Blust's
data, we found a different "top 10 most cohesive" pairs, as follows:
Table 8 Top 10 most cohesive pairs (Ross 2020: 368)
Pairing Cohesiveness Pairing (labels)

2b 4a 64% Cordilleran (Northern) GCP (CP)
4d 4e 60% GCP (Manobo) GCP (Danaw)
2a 2c 48% Cordilleran (Ilokano) Cordilleran (Central)
2a 4a 48% Cordilleran (Ilokano) GCP (CP)
4b 4e 47% GCP (South Mangyan) GCP (Danaw)
2b 4e 44% Cordilleran (Northern) GCP (Danaw)
4a 4d 44% GCP (CP) GCP (Manobo)
2c 2d 44% Cordilleran (Central) Cordilleran (Southern)
2a 2b 41% Cordilleran (Ilokano) Cordilleran (Northern)
3 4c 41% Central Luzon GCP (Palawanic)
The percentages calculated by Ross do not match the definition in his quote above.
For instance, 2b is shown to be affected by 425 innovations and 4a by 980 (the
figures I recalculated match here the ones in Blust 2020: 251), and 352 innovations
are common to 2b and 4a. If we divide 352 by 424+980-352 (352 is discounted to
eliminate double counting as shared innovations are both part of 424 and of 980),
one finds 33,5% (my calculation below), not 64% (Ross's calculation above).

Table 9 Top 10 most cohesive pairs (JFD's own computation)

2a 4a 44.2% Cordilleran (Ilokano) GCP (CP)
2b 4a 33.5% Cordilleran (Northern) GCP (CP)
2a 2d 27.3% Cordilleran (Ilokano) Cordilleran (Southern)
2d 4a 26.0% Cordilleran (Southern) GCP (CP)
2a 2b 24.7% Cordilleran (Ilokano) Cordilleran (Northern)
2c 4a 24.5% Cordilleran (Central) GCP (CP)
2c 2d 24.3% Cordilleran (Central) Cordilleran (Southern)
4d 4e 23.8% GCP (Manobo) GCP (Danaw)
2b 2d 22.1% Cordilleran (Northern) Cordilleran (Southern)
2a 2c 22.0% Cordilleran (Ilokano) Cordilleran (Central)
Still, six pairs are common to Ross's and my calculation, and his comments
(2020:368) are not significantly affected and remain true with my revised data: "it is
clear from these results that {2a, 2b, 2c, 2d and 4a} and [2b, 4a, 4b, 4d and 4e}
each form a cluster, and that 2b and 4a are in both clusters" (...) "the strength of the
relationship between 2b and 4a [jfd: in my calculation this pair remains the 2nd most
cohesive], and the overlap of the two most defined clusters suggests that at some
point in the (distant?) past, ancestors of the Cordilleran and Central Philippines
groups were adjacent (as the two microgroups are today".
On the other hand the calculations done by Ross probably follow more subtle
definitions that he does not share (but he gives references to theoretical work on
cohesion calculations on top of his page 368), I would take this point to a conclusion
for further research. Particularly the formula I have designed may give too much an
advantage to languages that have more thorough dictionaries. Still as the
conclusions drawn from my top 10s are not different from the one drawn from Ross's
top 10s, I will use the calculation I just specified for the rest of this short study.
3.2.2 Top groups clustering with South Mindanao

microgroups
We now look on the basis of the data I recalculated, which are the top microgroups
that are most cohesive with Gorontalic, Bilic, Sangiric and Minahasan.

3.2.2.1 Gorontalic (GCP)
Table 10 Decreasing cohesiveness of pairs involving Gorontalic

4g 7 13.3% GCP (Gorontalic) Sangiric
4g 8 10.6% GCP (Gorontalic) Minahasan
4d 4g 9.1% GCP (Manobo) GCP (Gorontalic)
4f 4g 8.8% GCP (Subanon) GCP (Gorontalic)
4c 4g 8.1% GCP (Palawanic) GCP (Gorontalic)
4e 4g 6.7% GCP (Danaw) GCP (Gorontalic)
4g 6 6.5% GCP (Gorontalic) Bilic
4g 5 6.0% GCP (Gorontalic) Kalamianic
4b 4g 5.9% GCP (South Mangya GCP (Gorontalic)
4a 4g 5.7% GCP (CP) GCP (Gorontalic)
2b 4g 5.0% Cordilleran (NortheGCP (Gorontalic)
2d 4g 4.8% Cordilleran (SoutheGCP (Gorontalic)
2c 4g 4.7% Cordilleran (CentraGCP (Gorontalic)
3 4g 4.5% Central Luzon GCP (Gorontalic)
(1 4g 4.3% Bashiic GCP (Gorontalic)
2a 4g 2.8% Cordilleran (Ilokano
GCP (Gorontalic)
Gorontalic has been shown to be part of the Greater Central Philippine subgroup,
but interestingly enough, it shows a greater number of "replacement innovations"
wtih Sangiric and Minahasan, two independent microgroups of PMP (and Bilic is not
far behind), than it does with members of its own (GCP) group. This may mean that
"replacement innovations" had started developing by contact between proto-GCP,
proto-Sangiric and proto-Minahasan (maybe even before the dialects of origin
crystallized into divergent proto-languages), and also that more innovations (those
not shared by Sangiric or Minahasan) developed within the proto-GCP language
itself, before again the latter split between proto-Manobo, proto-Subanon, proto-
Palawanic and proto-Danaw.
Not surprisingly, the cohesiveness with subgroups of the North (Cordilleran, Luzon,
Batanes) is very low, so the distant settlements must have occurred quite early after
the arrival of the Proto-MalayoPolynesian speakers.

3.2.2.2 Bilic
Table 11 Decreasing cohesiveness of pairs involving Bilic

4d 6 12.0% GCP (Manobo) Bilic
4e 6 11.0% GCP (Danaw) Bilic
4c 6 9.3% GCP (Palawanic) Bilic
4b 6 8.8% GCP (South Mangyan) Bilic
2c 6 8.3% Cordilleran (Central) Bilic
4f 6 8.1% GCP (Subanon) Bilic
2d 6 8.1% Cordilleran (Southern) Bilic
6 7 8.0% Bilic Sangiric
4g 6 6.5% GCP (Gorontalic) Bilic
4a 6 6.3% GCP (CP) Bilic
2b 6 6.0% Cordilleran (Northern) Bilic
2a 6 5.8% Cordilleran (Ilokano) Bilic
3 6 5.7% Central Luzon Bilic
5 6 5.3% Kalamianic Bilic
6 8 4.1% Bilic Minahasan
(1 6 3.2% Bashiic Bilic
Proto-Bilic and Proto-GCP had a fair amount of replacement innovations in common
before they "diverged by separation" - but much less than GCP shares with
Cordilleran - meaning that the Bilic people left the "Philippine" dialect cluster
significantly long before Cordilleran and GCP crystalized into two different proto-
languages. Surprisingly Bilic seems to have shared innovations with Sangiric almost
at the same level than with the Central and Southern Cordilleran; it might be
interesting to compute and compare the cohesiveness of a few triplets to see if any
more significant pattern would emerge.

3.2.2.3 Sangiric
Table 12 Decreasing cohesiveness of pairs involving Sangiric

4g 7 13.3% GCP (Gorontalic) Sangiric
7 8 12.3% Sangiric Minahasan
6 7 8.0% Bilic Sangiric
4e 7 7.5% GCP (Danaw) Sangiric
4f 7 7.1% GCP (Subanon) Sangiric
4d 7 6.3% GCP (Manobo) Sangiric
4c 7 5.9% GCP (Palawanic) Sangiric
5 7 4.6% Kalamianic Sangiric
4b 7 4.2% GCP (South Mangyan) Sangiric
4a 7 3.7% GCP (CP) Sangiric
2c 7 3.2% Cordilleran (Central) Sangiric
3 7 2.8% Central Luzon Sangiric
2a 7 2.2% Cordilleran (Ilokano) Sangiric
(1 7 2.2% Bashiic Sangiric
2b 7 2.2% Cordilleran (Northern) Sangiric
2d 7 1.1% Cordilleran (Southern) Sangiric
Sangiric people seem to have shared more innovations with geographically close
people such as Gorontalic, Minahasan and Bilic, than with any other Philippine
people, even the closest ones, "immediately beyond Bilic" (assuming the relative
positions were the same in the prehistory...) - the GCP Danaw, Manobo and even
(closer) Subanon. This may be evidence for a "South Philippine Cluster" (as Ross
seems to think there is) with all four people isolating themselves at the edge of the
"Philippine cluster" quite early in its history - or maybe that an important part of
Blust's data is still made of borrowings or "leakages" (if those shared vocabulary
items have been acquired after the 4 people migrated / separated).
Also let us remember the citation of Greenhill & al. 2010 already reproduced in
Section 1.3.2.3, that we repeat here for the reader's convenience:
"The Sangiric language subgroup is placed as a higher-order grouping within the
Western Malayo-Polynesian languages. The Sangiric languages are located in the
Sulawesi region but should be a primary branch of the Philippines family. Our
placement of Sangiric as a deeper group within the Western Malayo-Polynesian
linkage may either reflect contact-induced change with neighboring Sulawesi
languages, or it may reflect the repeated parallel drift that has occurred in
Sangiric and other Sulawesi-area languages" (p.6).

3.2.2.4 Minahasan
Table 13 Decreasing cohesiveness of pairs involving Minahasan

7 8 12.3% Sangiric Minahasan
4g 8 10.6% GCP (Gorontalic) Minahasan
4f 8 6.3% GCP (Subanon) Minahasan
4c 8 4.2% GCP (Palawanic) Minahasan
6 8 4.1% Bilic Minahasan
4e 8 3.7% GCP (Danaw) Minahasan
4d 8 3.4% GCP (Manobo) Minahasan
4a 8 2.9% GCP (CP) Minahasan
(1 8 2.7% Bashiic Minahasan
2b 8 2.4% Cordilleran (Northern) Minahasan
3 8 2.2% Central Luzon Minahasan
2a 8 2.1% Cordilleran (Ilokano) Minahasan
2d 8 1.9% Cordilleran (Southern) Minahasan
2c 8 1.9% Cordilleran (Central) Minahasan
5 8 1.5% Kalamianic Minahasan
4b 8 1.5% GCP (South Mangyan) Minahasan
This set is more surprising, only the top 3 seem to correlate what we might intuit of
the history of the Philippine x Sulawesi people. Other pairs are more surprising, such
as Palawanic and Bashiic - on the other hand they come very low in terms of
cohesiveness hence the order is probably not significant.
3.3 Conclusion
Of the 4 languages (microgroups in fact) that we just reviewed, Gorontalic, Sangiric
and Minahasan stay together and in the top of their 3 charts, they are closer to each
other (on the criterion set by Blust himself of looking at lexical replacement
innovation) than to any subgroup of the Philippines (as a geographic entity); apart
for Gorontalic that has been classified as a GCP language on phonological basis (and
actually other GCP members come immediately after Sangiric and Minahasan in the
Gorontalic cohesiveness table), it would be difficult to understand why Blust has
included them in his "resurrection" paper.
This is not the case for Bilic, that shows a decisively higher cohesion with GCP
languages of Mindanao.
4 Morphological innovations

In this chapter we will use as reference for comparison of verbal morphology
systems (limited to voice, tense, aspect, applicative affixes - excluding causatives,
distributive, reflexive etc...) the Tagalog paradigm shown by Reid (1992)' in his
paper "on the development of the aspect system of some Philippine languages", as
shown in the following extract (that Reid says is inspired from De Guzman 1978):
Figure 4 - Tagalog focus x aspect paradigms by Reid (1992:66)
Let us remember though that at least 2 features of this paradigm, according to Reid,
are innovations of "Central Philippines" within GCP - the usage of CV reduplication
to mark the imperfective, and the assimilation of -um- and -in- (+begun) where
they coexist. And "The only languages that are like Tagalog in using <in> to mark
a distinction between begun and not-begun verb forms are the other languages of
the Central Philippines..." (Reid 1992:74). The author indeed shows that for instance
Agutaynen (p.74) (a Kalamian language - micro-group in the Philippine cluster
descending directly from PMP, spoken in islands north-east to Palawan) - or Ilokano
(Northern Luzon)(p.69) both use <in> exclusively to mark '+completed'.
We will try to see how different languages of Mindanao may be from this
perspective. Also because our overall question in this paper is to find out evidence of
a past "Philippine" subgroup (be it issued from one single proto-language, or instead
from an already partly differentiated dialect cluster), we will also pick 2 or 3
languages spoken immediately south to Mindanao, to "test the boundary".

4.1 West-Coast Bajaw (Greater Barito / Sama-Bajaw / -)
We read Miller (2007)'s "Grammar of West-Coast Bajaw". West Coast Bajaw (ISO
639-2 'bdr' - henceforth "WC Bajaw") is one of the two Sama-Bajaw languages (9 in
total) not spoken in the Philippines (but in north-west Sabbah, Malaysia), and within
Sama-Bajaw is also one of the six "Sulu-Borneo" sub-sub-family members. Its closest
relative in the Philippines, part of the 3 "Borneo Coast Bajaw sub-sub-sub family"
members, is Mapun ('sjm').
The presentation by Miller of the voice system of WC Bajaw shows the following
resemblances and contrast with for instance Tagalog:
- the voice system seems to be a combination of the Philippine and the Indonesian
type - on the Philippine side, there is an alternance of an AV and UV focus, and
the suffix -an is presented as an applicative (with focus on the oblique argument
- the "instrument" applicative being marked with peN1- (p.79); we did not find
reference to any reflex of the *si- (i- in Tagalog) "instrument focus" though. The
author analyses the system as "symmetrical" rather than ergative, because even
though the patient focus is apparently unmarked (Ø affix) and more frequent,
both the AV patient and the UV actor remain core arguments
- the "true passive" is marked by -in- (which in Tagalog marks the "begun" aspect
and in other, non-CP Philippine languages, "accomplished"); there is no affix to
mark the tense and aspect, instead there are a 'completive particle' boi (p.329)
and a 'perfect aspect marker' ai (p.331)
We show below a copy of the voice affixes tables 3.1-2 pp.64-65.
Figure 5 West Coast Bajaw voice affixes by Miller (2007:64)

Figure 6 West Coast Bajaw intransitive affixes by Miller (2007:65)
Even though our evidence is not 100% solid as we did not find the paradigms for
those Sama-Bajaw languages spoken in the Philippines, we assume that the system is
similar for the whole Sama-Bajaw languages - which have been said by both Zorc
and Blust to not belong to their Philippine subgroup. Indeed this is then confirmed
by the verbal morphology, as expressed by Miller himself:
"Using the criteria established by Arka and Ross (2005) above, WC Bajau is an
Indonesian-type language. In WC Bajau there are two voices, actor and
undergoer. Since these voices are both syntactically transitive, WC Bajau has a
symmetrical voice system. In WC Bajau there is also a 'true' passive voice (in
which the actor argument is syntactically demoted), as is true of certain other
Indonesian-type languages such as Indonesian, Javanese, and Balinese (Ross
2002:458). WC Bajau has an applicative suffix (-an1) which promotes a variety
of oblique argument types to undergoer status, and this suffix can co-occur with
the actor voice prefix (N-). Thus WC Bajau meets the criteria established by Arka
And Ross (2005) for an Indonesian-type language." (Miller 2007:19).
4.2 Southern Subanen (GCP / Subanon / Eastern / -)

The Subanen (Philippine) languages are spoken in the Zamboanga peninsula of
Mindanao. Their classification and the reconstruction of proto-Subanon have been
done by Lobel (2013:303-358).
Table 14 Classification of Sobanon languages in Ethnologue
Language ISO2 02 03 04
Subanen, Central syb Greater Central Philippine Subanon Eastern
Subanen, Eastern sfe Greater Central Philippine Subanon Eastern
Subanen, Northern stb Greater Central Philippine Subanon Eastern
Subanen, Southern laa Greater Central Philippine Subanon Eastern
Subanon, Kolibugan skn Greater Central Philippine Subanon Eastern
Subanon, Western suc Greater Central Philippine Subanon
According to Lobel (2013:158), "Other languages [jfd: than Gorontalo-Mongondow
ones, see next section] in which rampant sound changes has taken place contain
even more complex inventories of allomorphs [jfd: of verbal affixes], the two most

extreme being Southern Subanen and Maranao, the latter of which combines rich
morphophonemic variation with a rich system of tense aspect marking similar to
that of Tagalog, Bikol, and Waray-Waray."
We show below a copy of Tables 5.6 and 5.7 (Lobel 2013:159) that detail Southern
Subanen verbal morphology.
Figure 7 *Mag-/*Pag- in Southern Subanen by Lobel (2013:159)
Figure 8 Central Subanen Verb Morphology by Lobel (2013:159)
In terms of differences from Tagalog, the most striking is that as there is no

"imperfective" aspect (shown in Tagalog by CV-reduplication), that <in> marks
"completed" ("PAST")(like non-CP Philippine languages then), that both the "non-
past" and "past" of the -um- AF keep their -um- affix (and the latter even persists
when -in- is inserted); conversely, the imperative loses -um-. An other point is that
there again, like Miller (2007) for West-Coast Bajaw, Lobel did not identify any
"instrumental" applicative (marked with i- in Tagalog).
4.3 Maranao (GCP / Danao / Maranao-Iranon / -)

Maranao is spoken in the two Lanao provinces of the Philippines, and in Sabbah in
Malaysia. Lobel (2013) discusses the Maranao paradigms at length under Tables 5.8

(p.161) and 9.8 (p.297), but we do not show them all here due to space constraints;
their main particularities seem to be 1. a usage of mag- affixes in the -um- verbs
paradigms, and 2. the Actor focus paradigms show a tremendous diversity of
phonological forms of the affixes, that also depend on the vowel of the first syllable
of the root.
We show though the following GCP recapitulation table of Object focus verb
conjugations in several GCP languages (including Maranao then). There again the
<in> has a different role than the one in Tagalog and other "Central Philippines"
(within GCP) languages - limited to "accomplished", rather than extended to "begun".
Figure 9 OF verb conjugations in GCP for Lobel (2013:298)
(Lobel 2013:298)
4.4 Mongondow (GCP / Gorontalo-Mongondow /

Mongondowic / -)
Blust (1991) has shown that the Gorontalo-Mongondow languages, spoken in
Sulawesi, Indonesia, are part of the GCP subgroup. They have been classified, and
proto-Gorontalo-Mongondow reconstructed, by Usup (1986).
Table 15 Classification of Mongondowic languages in Ethnologue
Language ISO2 02 03 04
Bintauna bne Greater Central Philippine Gorontalo-Mongondow Gorontalic
Bolango bld Greater Central Philippine Gorontalo-Mongondow Gorontalic
Buol blf Greater Central Philippine Gorontalo-Mongondow Gorontalic
Gorontalo gor Greater Central Philippine Gorontalo-Mongondow Gorontalic
Kaidipang kzp Greater Central Philippine Gorontalo-Mongondow Gorontalic
Lolak llq Greater Central Philippine Gorontalo-Mongondow Gorontalic
Suwawa swu Greater Central Philippine Gorontalo-Mongondow Gorontalic
Mongondow mog Greater Central Philippine Gorontalo-Mongondow Mongondowic
Ponosakan pns Greater Central Philippine Gorontalo-Mongondow Mongondowic
The interest for us here is limited, because as those languages are GCP, they cannot
help to validate or disprove the hypothesis of the past existence of a proto-Philippine
language - there is a fair consensus that at least at the lower classification level, GCP
had a proto-language, and hence is recorded as such in Ethnologue. We will just

look at the paradigms documented once again by Lobel (2013:157) for a high level
comparison with other paradigms already shown above:
Figure 10 Focus and Aspect in Mongondow for Lobel (2013:157)
Lobel observes (p.157) that "[it] exhibits more variation than most core Central
Philippine languages do." We see that again <in> marks the past, and that there is
no "contraction" of the combination <inum>. We also see that the "instrumental" i-
reappears.
4.5 Sangiric and Bilic missing...

As shown in Table xxx p.aaa above, two attested microgroups of Mindanao only are
not part of GCP - Bilic and Sangiric. Unfortunately xxx
4.5.1 Bilic
Bilic languages are spoken in the South of Mindanao. Their classification was done
by Savage (1986). We were able to procure this document but unfortunately it is
restricted to the phonological comparison / morphemes reconstruction of the proto-
language, with no data on verbal paradigms.
Table 16 Classification of Bilic languages in Ethnologue
Language ISO2 Where 01 02 03

Blaan, Koronadal bpr Philippines Malayo-Polynesian Bilic Blaan
Blaan, Sarangani bps Philippines Malayo-Polynesian Bilic Blaan
Tboli tbl Philippines Malayo-Polynesian Bilic Tboli
Giangan bgi Philippines Malayo-Polynesian Bilic
Tiruray tiy Philippines Malayo-Polynesian Bilic
4.5.2 Sangiric
Sangiric languages are in fact spoken mostly in North Sulawesi, Indonesia, but are
"at the boundary" of the Philippines, actually Sangil (or Sangirese; ISO 639-2 'snl') is

spoken in the Philippine Sarangani Islands at the South tip of Mindanao, and Talaud
and Sangir in Islands linking Sulawesi and Mindanao.
Their classification was done by Sneddon (1984). We were able to procure this
document but unfortunately it is restricted to the phonological comparison / -
morphemes reconstruction of the proto-language, with no data on verbal paradigms.
Table 17 Classification of Sangiric languages in Ethnologue
Language ISO2 Where 01 02 03

Sangil snl Philippines Malayo-Polynesian Sangiric Northern
Sangir sxn Indonesia Malayo-Polynesian Sangiric Northern
Talaud tld Indonesia Malayo-Polynesian Sangiric Northern
Bantik bnq Indonesia Malayo-Polynesian Sangiric Southern
Ratahan rth Indonesia Malayo-Polynesian Sangiric Southern
Note also the following quote from Sneddon (1984): "San [author's abbreviation for
Sangir] has been employed in a wider comparative study of Philippine languages by
Charles (1974). Charles believes, on lexical evidence, that San and Tontemboan - a
Minahasan language - lie outside the Philippine group. (...) Llamzon and Martin as
well as Walton place San and Snl [Sangil] within the Philippine language group but
both studies show them to be only distantly related to other Philippine languages.
Walton finds San and Snl to be a first-order subgroup of the Southern Philippine
group, one of the two first-order branches of the Philippine group." (p.4)
Actually as we mentioned that Sangiric is "at the frontier" between Philippines and
Sulawesi (Indonesia), let us complete with the observation that Sneddon (1984:11-
12) still speculates that Sangiric languages might be closely related to Minahasan
languages (5 languages of North Sulawesi, Indonesia, still today shown by
Ethnologue to constitute a separate microgroup of PMP, sibling of proto-Sangiric
then). If we had time and space we could look at this outer frontier as well...
4.6 Conclusion on the "verbal morphology" topic

It is a bit unfortunate that because of the Covid pandemic, the SIL library is not
accessible - we might have had access to more grammatical data on Sangiric and
Bilic, two microgroups of Mindanao that would be key to gauge whether
morphological information and verbal paradigms are more in line with phonological
evidence (only generally accepted method to attest the past existence of proto-
languages) or more in line with lexical evidence (that can also help identify
subgroups with a non-proto-language, but "dialect cluster" origin). Unfortunately,
most of the information we gathered about Mindanao verbal morphologies is from
GCP languages (and still we found interesting differences from the Central
Philippine smaller / dependent subgroup therein). We looked as a compensation to
languages of the Northern Borneo and even of the "non-Philippine" Bajaw languages

that do have representatives in the Philippines, and found indeed that the
differences were even more obvious. But time and space do not allow us to design
strict comparison measure criteria, so be it, at least we have identified sources that
could help us with further studies.
What is missing certainly are papers of the quality of the ones about "proto-
Austronesian verbal morphology" (Ross 2009) and "proto-Austronesian verb classes"
(Ross 2015), but for Malayo-Polynesian... opportunity for later research!
5 Conclusion
We started this review of subgroupings of the languages spoken in the Philippines
with two goals - try to find evidence of the past existence of a "proto-Philippine"
language, and give priority to data related to languages spoken in Mindanao.
We have found that the lexical evidence gathered by Zorc and Blust of an impressive
number of innovations claimed to be "exclusive" and "replacement innovations" is
compelling to accept that the Philippine languages before diverging into several
proto-languages, originate from a "Philippine subgroup" that many scholars,
following Ross (for Austronesian - starting with Oceanic) but also Schleicher, have
agreed to call "dialect cluster" or "linkage". On the other hand it seems that no
consensus has been reached so far to identify any phonological change that would
be specific and exclusive of the languages claimed to be "of the Philippine type" -
hence according to the generally accepted definitions of Comparative Linguistics, the
past existence of a proto-Philippine language has not been proven so far. Probably,
more work needs to be done to revisit the reconstructions that have been done of
proto-Malayo-Polynesian, as well as the ones of a putative proto-Philippine, before
they can be compared again - many linguists speculate that no difference will be
found between carefully reconstructed PMP and PPH.
In a last chapter we added a quick exploration of the paradigms of voice and aspect
of several of Mindanao languages, but the time lacks to lead a thorough comparison
of those and to draw any conclusion - we leave it open for later research to find out
where morphological evidence could stand aside phonological and lexical ones, a
question particularly interesting in the Philippines where the term "Philippine-type"
has been widely used to qualify the verbal inflection system. Would the apparent
similarities of this morphological data bear more, or less weight in the "proto-
Philippine existence" debate than the lexical one, is really phonological data the only
one that can be reliably used to attest the past existence of a proto-language?

List of Abbreviations and Symbols
(Alphabetical)
AN Austronesian PAN Proto-Austronesian

CEMP Central-East-Malayo- PCEMP Proto-Central-East-
Polynesian Malayo-Polynesian
CMP Central Malayo Polynesian
CP Central Philippines
EMP East-Malayo-Polynesian PEMP Proto-East-Malayo-
Polynesian
Ethnologue (see references)
F Formosan
GCP Greater Central Philippines PGCP Proto-Greater-Central-
Philippines
Glottolog (see references)
ISO International Standard
Organization
MP Malayo-Polynesian PMP Proto-Malayo-Polynesian
OC Oceanic POC Proto-Oceanic
PH Philippines PPH Proto-Philippines
RGH MP proto-phoneme in
Conant 1911's
reconstructions
RLD MP proto-phoneme in
Conant 1911's
reconstructions
SHWNG South Halmahera West New
Guinea
SIL Summer Institute of
Linguistic
WMP Western Malayo Polynesian

Bibliography
Arka, I. W., & Ross, M. (Eds.). (2005). The many faces of Austronesian voice systems—
Some new empirical studies. Research School of Pacific and Asian Studies,
Australian National University.
Blust 1974
Blust, R. (1983). More on the Position of the Languages of Eastern Indonesia. Oceanic
Linguistics, 22/23(1/2), 1–28.
Blust, R. (1989). Austronesian Etymologies: IV. Oceanic Linguistics, 28(2), 111–180.

https://doi.org/10.2307/3623057
Blust, R. (1991). The Greater Central Philippines Hypothesis. Oceanic Linguistics,

30(2), 73–129. https://doi.org/10.2307/3623084
Blust, R. (1993). Central and Central-Eastern Malayo-Polynesian. Oceanic Linguistics,

32(2), 241–293. JSTOR. https://doi.org/10.2307/3623195
Blust, R. (2005). A Note on the History of Genitive Marking in Austronesian Languages.

Oceanic Linguistics, 44(1), 215–222. JSTOR.
Blust, R. (2019). The Resurrection of Proto-Philippines. Oceanic Linguistics, 58, 153–

256.
Blust, R. (2020). Response to Comments on “The Resurrection of Proto-Philippines.”

Oceanic Linguistics, 59, 450–480.
Blust, R. A., & Trussel, S. (2015). Austronesian Comparative Dictionary, web edition:
Introduction. https://www.trussel2.com/ACD/intro-2016-02-25.pdf
Charles, M. (1974). Problems in the Reconstruction of Proto-Philippine Phonology and

the Subgrouping of the Philippine Languages. Oceanic Linguistics, 13(1/2), 457–
509. JSTOR. https://doi.org/10.2307/3622751
Conant, C. E. (1911). The RGH Law in Philippine Languages. Journal of the American
Oriental Society, 31(1), 70–85. JSTOR. https://doi.org/10.2307/3087491

De Guzman, V. P. (1978). Syntactic Derivation of Tagalog Verbs. University Press of
Hawaii. https://books.google.com.ph/books?id=CWkOAAAAYAAJ
Dyen, I. (1965). A Lexicostatistical Classification of the Austronesian Languages.

Waverly Press. https://books.google.fr/books?id=HYgOAAAAYAAJ
Ethnologue: refer to Lewis & al. 2016
Glottolog: refer to Hammarström & al. 2021.
Greenhill, S., Currie, T., & Gray, R. (2009). Does horizontal transmission invalidate
cultural phylogenies? Proceedings. Biological Sciences / The Royal Society,
276, 2299–2306. https://doi.org/10.1098/rspb.2008.1944
Greenhill, S., Drummond, A., & Gray, R. (2010). How Accurate and Robust Are the
Phylogenetic Estimates of Austronesian Language Relationships? PloS One, 5,
e9573. https://doi.org/10.1371/journal.pone.0009573
Greenhill, S. J., & Gray, R. D. (2009). Austronesian language phylogenies: Myths and
misconceptions about Bayesian computational methods. Austronesian Historical
Linguistics and Culture History: A Festschrift for Robert Blust. Pacific
Linguistics, Canberra, 375–397.
Hammarström, H., Forkel, R., Haspelmath, M. & Bank, S.. 2017. Glottolog 4.5.
Leipzig: Max Planck Institute for Evolutionary Anthropology.
https://doi.org/10.5281/zenodo.5772642 (Available online at
http://glottolog.org), Accessed on 2022-01-07.
Lewis, M. P., Simons, G. F., & Fennig, C. D. (2016). Ethnologue: Languages of the
World, Nineteenth edition. SIL International. http://www.ethnologue.com
Liao, H. (2020). A Reply to Blust “The Resurrection of Proto-Philippines.” Oceanic

Linguistics, 59(1–2), 426–449.
Llamzon, T. A. (1975). Proto-Philippine Phonology. Archipel, volume 9, 1975.

https://doi.org/10.3406/arch.1975.1214
Lobel, J. W. (2013). Philippine and North Bornean Languages: Issues in Description,

Subgrouping, and Reconstruction [PhD. Dissertation]. University of Hawai’i at
Manoa.

Miller, Mark T. (2007). A Grammar of West Coast Bajau. University of Texas at
Arlington.
Pawley, A., & Green, R. C. (1984). The Proto-Oceanic Language Community. The
Journal of Pacific History, 19(3), 123–146.
Pawley, A. (2006). Origins of the Filipinos: Linguistic and archaeological evidence.
Paz, Consuelo. J. (1981). A reconstruction of proto-Philippine phonemes and morphemes.

The Cecilio Lopez Archives of Philippines Languages & The Philippine
Linguistic Circle.
Reid, L. A. (1982). The Demise of Proto-Philippines. In A. Halim, L. Carrington, & S. A.

Wurm (Eds.), Papers from the Third International Conference on
Austronesian Linguistics (pp. 201–216). Pacific Linguistics.
Reid, L. A. (1992). On the Development of the Aspect System in Some Philippine

Languages. Oceanic Linguistics, 31(1), 65–91.
https://doi.org/10.2307/3622966
Reid, L. A. (2007). Historical Linguistics and Philippine Hunter-Gatherers.

http://scholarspace.manoa.hawaii.edu/handle/10125/33036
Reid, L. A. (2017). Revisiting the position of Philippine Languages in the Austronesian

Family. The Br Andrew Gonzalez FSC (BAG) Distinguished Professorial Chair
Lecture, De La Salle University, Manila.
Reid, L. A. (2018). 6. Modeling the Linguistic Situation in the Philippines. Senri

Ethnological Studies, 98, 91–105.
Reid, L. A. (2020). Response to Blust “The Resurrection of Proto-Philippines.” Oceanic

Linguistics, 59(1–2), 374–393.
Ross, M. (1988). Proto Oceanic and the Austronesian Languages of Western

Melanesia. Department of Linguistics, Research School of Pacific Studies, the
Australian National University.
https://books.google.fr/books?id=eFdxAAAAIAAJ
Ross, M. D. (1994). 3. Some current issues in Austronesian linguistics (D. T. Tryon, Ed.;
pp. 45–120). De Gruyter Mouton.
https://doi.org/10.1515/9783110884012.1.45

Ross, M. (2002). The history and transitivity of western Austronesian voice and voice
marking. In F. Wouk & M. Ross (Eds.), The History and Typology of Western
Austronesian Voice Systems (pp. 17–63). Research School of Pacific and
Asian Studies, ANU.
Ross, M. (2005). The Batanic Languages in Relation to the Early History of the Malayo-
Polynesian Subgroup of Austronesian.
Ross, M. D. (2020). Comment on Blust “The Resurrection of Proto-Philippines.” Oceanic

Linguistics, 59(1–2), 366–373.
Savage, T. D. (1986). A reconstruction of Proto-Southern Mindanaon. Studies in

Philippine Linguistics, 6(2), 181–223.
Smith, A. (2017). The Western Malayo-Polynesian Problem. Oceanic Linguistics, 56,

435–490. https://doi.org/10.1353/ol.2017.0021
Sneddon, J. N. (1984). Proto-Sangiric and the Sangiric Languages. Department of

Linguistics, Research School Pacific Studies, Australian National University.
https://books.google.fr/books?id=M1oOAAAAYAAJ
Tryon, D. T. (2006). Proto-Austronesian and the Major Austronesian Subgroups. In

The Austronesians: Historical and Comparative Perspectives (pp. 19–41).
Zorc, R. D. (1986). The genetic relationships of Philippine languages. Pacific Linguistics

C, 94(FOCAL II: papers from the Fourth International Conference on
Austronesian Linguistics), 147–173. https://doi.org/10.15144/PL-C94.147
Zorc, R. D. (2020). Reactions to Blust “The Resurrection of Proto-Philippines.” Oceanic

Linguistics, 59(1–2), 394–425.

Western Malayo Polynesian Languages Micr

Uploaded by

Copyright:

Available Formats

Western Malayo Polynesian Languages Micr

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Western Malayo Polynesian Languages Micr

Uploaded by

Copyright:

Available Formats

3

Western Malayo-Polynesian Languages

presented to Dr. Irma Peneyra

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 i

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 ii

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 1

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 1

1.1 Purpose and content

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 1

To give a reasonable goal to this review, we steer it towards looking at data

We will start (in Section 1.2) by an introduction to the languages spoken in

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 2

Table 1 Philippines language microgroups

"Philippine" subfamilies Where

We also show in Table 2 the subgrouping of the Malayo-Polynesian languages not

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 3

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 4

non-Philippine MP languages: 1060

Table 3 Regions where Philippine MP languages are spoken...

Philippine Malayo-Polynesian 176 Luzon Visayas Mindanao Subfamilies? (Other comments)

Table 4 Regions where Greater Central Philippine Languages are spoken...

Greater Central Philippine 90 Luzon Visayas Mindanao Subfamilies? (Other comments)

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 5

Table 5 Regions where Central Philippine languages are spoken

Central Philippine 52 Luzon Visayas Mindanao Subfamilies? (Other comments)

Table 6 Malayo-Polynesian languages microgroups spoken in Mindanao

Philippine Malayo-Polynesian 176 Luzon Visayas Mindanao Subfamilies? (Other comments)

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 6

Philippine Malayo-Polynesian 176 Luzon Visayas Mindanao Classification

1.3 Methodology: two kinds of subgroups

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 7

1.3.1 AN wave theory: Ross (1994) on "linkages"

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 8

Figure 1 Austronesian Family Tree by Blust (2009:2)

Particularly and for example, the MP "Proto-Malayo-Polynesian" node has two

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 9

1.3.2 From lexicostatistics to the ACD to... lexicostatistics

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 10

1.3.2.1 Lexicostatistics - and its application to the Philippines

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 11

1.3.2.2 Comparative word lists and dictionaries

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 12

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 13

1.3.2.3 Bayesian Phylolinguistics

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 14

Figure 2 Bayesian Tree, Philippine portion (Greenhill & Gray 2009:12)

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 15

1.3.3 The "resurrection of proto-Philippine" debate

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 16

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 17

1.3.4 Blust on the importance of phonological evidence

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 18

2.1 Phonological innovations: literature review

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 19

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 20

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 21

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 22

Lingg385 – Delmer - WMP micro-groups of Mindanao - January 9, 2022 23

2.2 Phonological innovations across and within PH

2.2.1 The RGH Law

2.2.2 The merger of d and z

2.2.2.3 Discussion and conclusion on the d/z merger

Figure 7 Mag-/Pag- in Southern Subanen by Lobel (2013:159)