UNDERSTANDING TAGALOG VERBS: Part 3: CV-reduplication in the Incompleted Aspect1 Resty M. Cena Tagalog indicates the incompleted action in the verb by reduplicating a CV of the stem. For example: bi-bili, magbi-bili, magpa-pabili, maki-kibili, etc.
[bili “buy, sell”]
CVr is the second part of the m-...-CVr time morpheme in the language, where mrepresents [+/-Begun] and CVr [+/-Completed]. m- (with allomorphs m and n) are purely tense markers. In um-/-in- verbs, they migrate inside the root, possibly in accordance with phonotactic rising sonority. m-takbo n-takbo m-agtakbo
> > >
t-m-akbo > t-n-akbo >
t-um-akbo t-in-akbo m-agtakbo
This phenomenon has been discussed before (see relevant studies below). In this note, I present some observations. CVR ASPECT REDUPLICATION IN TAGALOG The CV after the first ‘strong’ morpheme boundary is a candidate for CVr. There are two modifiers to the rule: (i) Only one application of CVr in a derivation2 (there’s an exception), and (ii) There is a tendency to apply CVr to the CV closest to m-. The strong morpheme boundaries are the word boundary (marked by # below), and the internal morpheme boundary (marked by + below) that separates the mode affixes {-a-, -ag-, -ang-} from the rest of the verb form. Throughout, I will refer to the affixes {-a-, -ag-, -ang-, pa(ti)-, ka, ki-, kang(da)-, -sa-, -si-} as mode affixes (that is to say, adverbialish affixes that indicate the mode or manner of the action). Mode affixes are independent of aspect and voice affixes. 1 Translated from Morpolohya ng Tagalog (Ceña 2017, real soon now)). 2 The recent perfective form, for example, kabibili, does not undergo CVr; my guess is that the lexical CVr of the root ka-bi-bili prevents another one (see rider (i) on the rule above).
1
#bili m-a+bili m-ag+bili m-ang+gapas
> > > >
#bi-bili m-a+bi-bili m-ag+bi-bili m-ang+ga-gapas
In the examples below, the null symbol “ ø “represents the [-Begun] “contemplated” form of m- in the basic -um-/-in (bumili/binili) paradigm. m-CVr#bili > m-CVr#padalhan >
ø-CVr-bili ø-CVr-padalhan
> >
bi-bili pa-padalhan3
[dala “to carry”]
Where the prefix is a mode affix with an onset, for example, pa-dalhan, CVr may apply on the affix (i below), where m-CVr respects the word boundary and thus applies from the outside, or on the root (ii below) where CVr obeys rather the morpheme boundary. The infix -in- is the [+Begun] form of m-, again in -um-/-in verb type. m-CVr-padalhan
(i) m-pa#padalhan (ii) pa+da-dalhan
> >
p-in-apadalhan p-in-ada-dalhan
These mode affixes {-ka-, -ki-, -sa-, si-} don’t trigger CVr: ma-katapos ma-kibili mag-sidala mag-saulo
?maka-tatapos *maki-bibili *magsi-dadala *magsa-uulo
Preferred CVr ma-kakatapos ma-kakabili mag-sisidala mag-sasaulo
[tapos “to finish’]
[saulo “to memorize”]
Exceptionally, the affixes -ti- in pati- and -da- in kangda- do not reduplicate,4 despite them always following a strong morpheme in (p)a- or (k)ang-, as shown below. Videa de Guzman (p.c.) considers pati- as a single morpheme, and kangda- may be similarly treated. They are non-monosyllabic mode affixes. magpa+tihulog magkang+dahulog
*magpa-titihulog *magkang-dadahulog
Compare mag-papatihulog magka-kangdahulog
CVr cannot reduplicate its other half, m-. CVr applies before m- merges, and hence it is unavailable for CVr, among other reasons. For example: 3
4
Pa-padalhan is slightly odder than pada-dalhan. In magkandadapa, > magkandarapa, da- is not CVr but is a part of kangda-.
2
mabili > magdala > manguha >
*ma-mabili *ma-magdala *ma-manguha
[kuha “to get”]
These forms are not to be confused with the forms in the third column below, where the underlined ma- is not tense, but part of the stem. That ma- is not tense is evidenced by the fact that the tense affix -in- can be inserted into ma-, this m-in-abuti, min-amabuti. Adjective mabuti magdamag
> >
Verbalized mabutih-in magdamagin
> >
CVr reduplicated ma-mabutih-in [buti “ to make good”] ma-magdamag-in [damag “overnight”]
There is a preference to reduplicate the CV of the first mode affix (i.e., the stem) rather than the root. ma+pag+tapos ma+pang+kuha mag+pa+tapos
CVr on stem ma-papagtapos ma-papang(k)uha mag-papatapos
CVr on root mapag-tatapos *mapa-ngunguha [nguha < ng-kuha] *magpa-tatapos
Notice that the alternative forms below are due to the opportunities provided by the two + morpheme boundaries in each word. Again, the preference is to reduplicate the CV closer to m-. With three-syllable prefixes, the first CV of the stem (i below) or of the root (iii) may undergo CVr. The acceptability of (i) is expected, the involved CVs being closest to m-, but the well-formedness of (iii) is somewhat of a surprise, in light of the fact that the CV in (ii) is closer to m- than that of the CV in (iii) to its m-. Perhaps the rule is to reduplicate the first CV of the stem or the root. Affixes ma+kapagma+kipagma+kapangma+kipangma+pag+pa ma+pag+pang
(i) ma+kakapag+dala ma+kikipag+tanim ma+kakapang+(k)uha ma+kikipam+ili ?ma+papagpadala ?ma+papagpang(k)uha
(ii) *ma+kapapag+dala *ma+kipapag+tanim *ma+kapapang+(k)uha *ma+kipapam+(b)ili ?ma+pa+papagdala ?ma+pag+papang(k)uha
3
(iii) ma+kapag+dadala ?ma+kipag+tatanim ma+kapang+(ng)unguha ma+kipam+(m)imili ma+papag+dadala *ma+pag+pang(k)unguha
But with prefixes of more than four syllables or more,5 the root becomes too distant to pull in CVr to its side. makapagpa-
makapagpang-
magpakipa-
magpakipagpa-
makapakipang-
mangagsipag-
ma-kakapagpakuha [kuha “to get”] ?maka-papagpakuha makapag-papakuha *makapagpa-kukuha ma-kakapagpangita [-ngita < kita “to meet up”] ?maka-papagpangita makapag-papangita *makapagpang-ngingita mag-papakipatanim [tanim “to plant”] *magpa-kikipatanim *magpaki-papatanim *magpakipa-tatanim mag-papakipagpalakihan [laki “big, large”] *magpa-kikipagpalakihan *magpaki-papagpalakihan ??nagpakipag-papalakihan *magpakipagpa-lalakihan ma-kakapakipanggapas [gapas “harvest rice”] ?maka-papakipanggapas *makapa-kikipanggapas *makapaki-papanggapas *makapakipang-gagapas ma-ngangagsipagdala [dala “to carry”] mangag-sisipagdala *mangagsi-papagdala ?mangagsipag-dadala
Why is there a preference for CVr to apply on the CV closest to m-? It serves some purpose for parts of a morpheme to stay close together, it seems. And the exception? CVr in forms with the mode affixes -agkang- applies more than once in a derivation. mag+kang+
mag-kakang-huhulog mag-kakang-luluha mag-kakang-susuka mag-kakang-hihilo
5 This is where the native speakers’ intuition will be taxed to the max. Let me know if you disagree with me on my judgment on these forms to the extent that my observations will be without basis. 4
mag-kakang-uutot
Similar to -kang-, two pseudo ‘major morphemes’ appear to be excemptions as well. One’s the dead affix -hing-, which, like a zombie, refuses to rest, and the other is sing-, an import from adjective morphology. They reduplicate, as in (i) below, but, unlike -kang-, do not themselves trigger CVr, thus: (i) mang-hing-kuto mag-sing-laki
(ii)
> mang-hihinguto > mag-sisinglaki (v.)
*manghing-kukuto > *maghing-kukuuto *magsing-lalaki
And, lastly, a piece of data that is more of a curiosity than an exception. The voice affix `i- undergoes and triggers CVr.6
maibili
m-CVr+`i+bili >
(i) maiibili
(ii) maibibili
?ma-`i`ibili
ma-`i-bibili
By itself, it does not reduplicate, thus: `i-`ibili, and thus by our assumption it is preceded by a strong morpheme boundary (compare: #pabilihin > #pa-pabilhin). It can be preceded by and only by the mode affix -a-, and it is this strong morpheme that allows ito redupliate, as in (i) above. Thereupon the tense m- attaches to -a-, completing the derivation. But i- truly is an interesting case. Of the three overt voice affixes (the other two are -in and -an), i- is the only one that occurs in pre-root position (for example, as in i-bili, i-kabili, i-pabili, i-saulo). This suggests that it migrated to its surface position, to the left of m-CVr, as shown in (i) below; this is the only way to derive (ii) below within the assumptions laid in this note. `i- [m-CVr-pabili
-i ]
>
`i-p-in-a-bi-bili
Failure to insert m-CVr in the derivation in its position above allows “another round” of derivation, although only the mode affix -a- may be attached to i-, and after which m-CVr is inserted to the left of -a-, thus: m-CVr-`i-pa-bili
Of the three overt voice affixes (the other two are -in and -an), i- is the only one that occurs in preroot position (for example, as in i-bili, i-kabili, i-pabili, i-saulo). 6
5
Thereupon, there opens a fresh round of CVr: ma`i`ipabili, ma`ipapabili, ma`ipabibili ma`ipakikibili, ma`i`ipakibili ma {-`ipagpa-, `ipagpang-, -`ipakipagpang-, `ikakapagpakipang-... } etc.
Exceptions aside, the data show that incompleted aspect reduplication applies on the closest CV to m-CVr, and, alternatively, on the root, when the root is not too far away, in linear distance. Relevant Studies
(I am unable to access Carrer’s and Travis’ works. I think that a no-review of the literature is more honourable than one that neglects other relevant studies.) Carrier, Jill (1979). The Interaction of Phonological and Morphological Rules in Tagalog. Doctoral Dissertation, MIT. Rackowski, Andrea (1999). Morphological Optionality in Tagalog Aspectual Reduplication. In Karlos Arregi, Benjamin Bruening, Cornelia Krause and VivianLin, eds., Papers on Morphology and Syntax, Cycle Two, MITWPL 34: 107-136. Cambridge: MITWPL. Samuels, Bridget. (2006). Reduplication & Verbal Morphology in Tagalog. Harvard Linguistic Theory Group. Travis, Lisa. Inner Aspect. Dordrecht: Kluwer Academic Press.
6