Archipel

Numeral expressions in Tagalog

Jean-Paul Potet

Citer ce document / Cite this document :

Potet Jean-Paul. Numeral expressions in Tagalog. In: Archipel, volume 44, 1992. pp. 167-181;

doi : https://doi.org/10.3406/arch.1992.2860

https://www.persee.fr/doc/arch_0044-8613_1992_num_44_1_2860

Fichier pdf généré le 21/04/2018

Jean-Paul G. POTET

Numeral Expressions in Tagalog

Introduction
Among the curiosities of the old Philippine culture, the numeral system
stands apart as a decimal system. The Tagalog numeral system as we know
it nowadays is a rather modern development. It was unknown when the
Spaniards conquered the archipelago (Cebu 1565-Manila 1571). Instead,
contemporary Tagalogs used complex numeral expression which, though duly
recorded by those worthy missionary linguists of yestercenturies, seem to have
passed unnoticed by historians of mathematics. I hope this monograph will
draw their attention to this oddity, and that they will find out why it came
to exist as it was.
The vocabulary, common to both systems, is presented in the first part,
the modern system described in the second part, and the old one analysed
in the third part.

Preliminary remarks
Before I proceed any further, let us define some key terms. A number is
an abstract quantifier expresed either through a numerical or a numeral
expression. A numerical expression is the digital representation of a number
whereas a numeral expression is its verbal representation. In simple terms,
a numeral expression is the reading of the corresponding numerical
expression. Numerical expressions vary according to the base (or radix) chosen. The
base of the decimal system is ten. That of the binary system is two. For
example, the number represented by 5 in the base-10 system will be represented
by 101 in the binary system. Whatever the digits used, these numerical systems
are universal. Such is not the case with numeral expressions, which vary from
one language to another.
Every number is the sum of a series of products, and the multiplicand of
the last product is 1, e.g. 96,185 = 9.104 + 6.103 + 1.102 + 8.10 + 5.1. Each
product belongs to an order. Orders are ranked from the lowest to the hight-
168

est. Thus 5.1 belongs to the first order, and 9.104 to the sixth order. In the
decimal system, ten units of any order make up one unit of the order above.

Conventions
Apart from some quotations from old books left in the original spelling,
all examples are respelt in my spelling system which is a compromise
between that of the Institute of National Langaage/Surian ng Wikang Pambansa
and a broad phonetic transcription. I use the letter q to represent the final
glottal stop. Initial and medial glottal stops being taken for granted are not
recorded. Sup- and sub-script letters stand for phonemes that are not
represented in conventional spelling. A supscript phoneme is realized as a phone.
A subscript phoneme is only virtual, therefore silent. The ligature w (L) is
represented by 'ng. Readers who wish to see examples as they are actually
spelt may just ignore all the diacritics.
Owing to some limitations in the types available with Archipel, the empty-
set symbol has been replaced by underlined 0 (0^ . Spanish n with tilde by
underlined n (n), and the eng phonetic sign by ng.

1. Vocabulary
The vocabulary of numeral expressions falls into two sets : multipliers and
multiplicands.

1.1. Multipliers
The multipliers are the integers from 1 to 9.
1 isâh 6 ânim
2 dalawâh 7 pitôh
3 tatlôh 8 walôh
4 âpat 9 siyâm
5 limâh

These are dissyllabic radicals, except dalawâh and tatlâ1, which result
from the prefixation of the duplicand of the radical's initial syllabe.
The radical of dalawah is dawâh: da- + dawâh = dalawdK The làl of dawâh
became [1] because it was intervocalic, as happened sometimes in old Taga-
log. In middle and modern Tagalog, intervocalic /d/ is realized as [r] in a
majority of cases, and remains [d] in a minority of case, e.g. dadn (hundred), lahft
sa raân (more than hundred, see infra).
TatW is a syncopated form; its radical is talfi : ta- + talâ1 = *tatalâi =
tatlfr.
According to Dempwolff (1934-1938), limâh 'five' also meant 'hand' in
Proto-Austronesian - along with *lingah -, but this is not true in Tagalog
in which 'hand' is said kamây.
The radicals dawcP1, tald1, dpat (cf. Romanian patro) and pita1 sound Indo-
european. Like apôy (fire), *ig in tubig (water, cf. irrigation) and other such
items, these might be evidence of early contacts between Austric and Indo-
European tribes.
Isah is often reduced to ^a^.
 169

1.2. Multiplicands
The multiplicands are the powers of ten.

10 pu<*
102
103 daân (thousand)
libôh (hundred) 106 âtih
107
108 ângaw
(hundred
(ten(million)
million)
million/2)
103 laksâ<i (ten thousand) 109 gatôs (billion/3)
105 ytita^ (hundred
thousand)

is the contraction of puô% Puô% in its turn, comes from pule?1. Even
in the old system puld* was only used when quoted apart. In numeral
expression it was always replace by puô\ Pulo11 became polo in Spanish and was
used to refer to a tax levied on the natives and paid as labour, goods and/or
money. Actually, the right word was ikapuloh (tenth) as evidenced by an
undated Church pamphlet on the faithful's duties.
(1) At sa icalimang utos manga bilin
ang Dios ay atin na papamaguhin
sa lahat nang atin na manga pananim
at ang icapulo'i sa Dios ihayin.
(Casaysayan nang Abecedario, 1926 edition, p. 20)
And God's fifth commandment is that we give to God the tithe of our first crop
of what we have planted.
Some of these radicals are borrowed from Sanskrit with a modification
of their value. Compare Sans, laksa (hundred thousand) and Tag. laksâ* (ten
thousand), Sans, ayuta (ten thousand) and Tag. yûtefi (hundred thousand).
Only Sans, koti (ten million) and Tag. kâtfr (ten million) tally. Laksa appears
in Coedès (1930 : 76).
Today's Tagalog speakers do not know the multiplicands beyond liboh
(103). Instead they use milyôn (106) and bilyôn (109). Schachter and Otanes
(1972 : 200) only mention laks& and ângaw as possibly used by a minority
of speakers. So does Llamzon (1976 : 118), who qualifies such words as bookish.
My own collection was collated from texts and dictionaries.

The modern system

In the modern system every product is translated according to the
structure m + L + M (multiplier + ligature + multiplicand), M being a power of
ten, e.g. âpat na daân (four + L + hundred) 'four hundred'.

2.1. Use of the ligature

When the ligature has the form -ng, it is often spelt according to its
pronunciation; it becomes [m] before labials, [n] before palatals, and [ng] before
velars. It remains [ng] before glottals and glides. Let it be recalled that the
ligature is the leftmost item of a given group (Dell 1981).
(2)Tatl6h + [L + pu<i]'= tatluh'ng pui = tatlumpui
three + [L + ten] = thirty
170

(3) tatlôh + [L + daân] = tatlûh'ng daân = tatlundaân

three + [L + hundred] = three hundred

(4) tatlôh + [L + ângaw] = tatlûh'ng ângaw = tatlûng-ângaw

three + [L + million] = three million.

2.2. Counted item

If the counted item is mentioned, the noun that stands for it is connected
to the numeral expression by the ligature. Besides, the plural marker is not
used.
(5) tatlumpû^ + [L + aklât] = tatlumpûq'ng aklât
thirty + [L + book] = thirty books.
If the plural marker is added before the numeral expression, the latter
becomes approximate.
(6) mangâh + tatlumpûq'ng aklât = mangâh tatlumpûq'ng aklât
pi. + thirty books = about thirty books.

2.3. First-order numbers

A first-order number is obtained by using the unit one as a multiplicand
without giving it any phonetic form. Owing to the fact that the ligature goes
with the unit, it will disappear with it. To make things clear for the reader,
I shall use the symbol I for the first order. If the number thus obtained is
used to count things, the number and the noun standing for the thing
counted are connected by the ligature.
(7) dalawâh + [L + I] = dalawâ/ng I) = dalawâh
two + [L + I] = two

(8) dalaâh + [L + I] + [L + aklât] = dalawâ^ng I)'ng aklât = dalawâ^ng aklât

two + [L + I] + [L + book] = two books.

2.Jf. Order connection

If several orders are involved, each m-L-M group is linked to the
preceding one by the connector at (and), which is often shortened to 't. Like the
ligature, this connector is the leftmost item of its group, hence the general
formula : m-L-M + n{at + m-L-M).
(9) [âpatnaraâj't dalawâj'ng aklât]
[four hundred] + [and + thirty] + [and + two] + [L + book]
four hundred and thirty two books.
This structure with at (and) between the products is a caique of the old
Spanish structure in which y (and) played the same part, as contrasted to the
current Spanish structure in which y is used only between tens and units.
 171

(10) an o de mil y quinientos y setenta y uno

year ôî thousand and five-hundred and seventy and one
(Morga 1609 [Rizal 1890] :10)
taôn'ng isâ^'nlibOjj't limâjj'ngdaân at pituh'mpû 't isâh
year-L one-L-thousand-and five-L-hundred and seven-L-ten and one
(1964 translation :7)
the year fifteen seventy-one.
This was made possible by the existence of the multiplicative structure
«mLM ay result ».
(11) Âpat na limâh ay dalawâ^mpû1
four L five [anteposer (4)] two-L-ten
Four times five equals twenty.
To achieve the modern system, all that Tagalog speakers had to do was
to represent each product by the m-L-M part on the left of the anteposer.

2.5. The second order

The series from eleven to nineteen is a remnant from the old system. We
shall see its formula in the following section.
11 labfh'ng-isâh 16 kabîh'ng-ânim
12 labîh'ndalawâh 17 labfh'mpitôh
13 labfh'ntatlôh 18 labfh'ngwalôh
14 labfh'ng-âpat 19 labfh'nsiyâm
15 labfh'nlimâh

2.6. Ordinals and fractionals

Ordinals are obtained by prefixing ika- to a cardinal expression, e.g. ika-
dalawd^mpTSfl « twentieth ». In the case of « second » and « third », the base
is the radical not the numeral itself : ikalawâh « second », ikatlâ1 « third ».
The word for « first », ûnah (naûnah in old Tagalog), is the only exception.
It is not a borrowing from Spanish where « first » is expressed by primero
(masc.) and primera (fern.). Mûna)1 (firstly), is derived from this radical : ma-
+ ûnah = maûnah = mûnah.
Apart from kalahâti* /half), fractionals are obtained by prefixing sa- to
ordinals, but the current trend is to drop saika-, e.g. dalawa\ ng saikatloh I
dalawa\ng katlfr « two thirds ».
There is another ordinal series in pang-. Pang- is directly prefixed to
ûnah : pangûnaK A prefix ka- is added to the radicals for « two » and
« three » the initial of which is absorbed by the final of pang- : pang- + ka-
+ lawah = pangalawah (second); pang- + ka- + talfr = pangatlâ1 (third).
Beyond thids point, ka- is erased, e.g. *pang- + ka- + limâh = panglimâh
(fifth).

2.7. The present situation

Though native speakers now tend to use Spanish and English substitutes
(pronounced the Filipino way) in many instances due to either coercive situa-
172

tions (e.g. telling the time, attending a maths lesson) or the speaker's own
style, the system is still alive.
As a computing system, however, it is on its way out. Experience proves
that many Tagalogs find it difficult to do any of the four arithmetical
operations in their native language. The majority will just reckon in English and
translate the resulting figure into Tagalog.

3. The old system.

3.1. Sources
The old system is described in San Joseph (5) (1610 [1752] : chapter 19,
664-698), Totanes (1745 [1796] : 118-120]) and Laktaw (1929 [posthumous]) :
351-356). San Joseph, the best source for old Tagalog, did not mention any
other system. About a century and a half later, Totanes' s description marked
no change, but he explained that the natives would use the modern system,
when addressing Spaniards.
(12) Aunque yà con la communication de los Esparôles, muchos cuentan como nosos-
tros, y asi dicen : dalavàng pôvo at ysà, veinte y uno. sàngdaan at lima, ciento
y cinco. limàng daan dalavàng pôvo at lima, quinientos y veinte y cinco, y asi de
los demas numéros.
(Totanes 1796 : 120)
Though today, when communicating with Spaniards, many are those who count
as we do. Thus they will say : « dalawà\'ng puô* at isâh », twenty one; « sandadn
at lim<$ », one hundred and five, « limâ^ 'ng dadn'ng dalawà\'ng puô* at limdh »,
five hundred and twenty five, and so on with the other numbers.
It is extremely difficult to determine the period during which the
Tagalogs switched from the old to the modern system because examples in
contemporary documents are few and far between. For instance the date at the
bottom of a 1665 petition (Potet 1987) is in Spanish. Conversely that of May
26, 1681 at the bottom of an unpublished petition from Mayobog aldermen
(Archivo Franciscano Ibero-Oriental, legajo 094/18) is in Tagalog, and, partly,
in the old system.
(13) labi sa libon anim na daan valong pouo at ysang taon
more than thousand-L six L hundred-^ eight-L ten and one-L year
sixteen eighty-one (1681).
Laktaw presented this system as « (el) antiguo modo de contar de los
Tagalogs » (The old way of counting of the Tagalogs), and specified that it was used
until the beginning of the Nineteenth Century (hasta los comienzos del siglo
XIX) . Because it had passed off long before his generation, he felt it
necessary to add an appendix on the old system so that his readers should be able
to understand old documents. He also expressed the opinion that the modern
system was due to the influence of the Spanish civilization.
As a surviving exponent of Nineteenth-Century philology, Marre (1902)
described the old numeral system in his remarkable grammatical pamphlet.
Even though he quotes no source, it is pretty obvious that his information
came from Spanish Tagalog grammars since he sounds as if he believed that
the old system was still in use at end of the Nineteenth Century. His exam-
 173

pies are correct, but the way he presents them suggests that he did not
understand how the system worked.
So far, in the Twentieth Century proper, no philippinologist seems to have
been aware of its existence. The failure of modern linguists to link up with
their worthy predecessors is striking, and can be easily accounted for by their
general disregard for traditional scholarship. In so far as Bloomfield (1917),
Schachter and Otanes (1972), Llamzon (1976) and all other Twentiech-Century
authors do not mention it, even in a passing way, it is pretty obvious that
they had never heard of it when they wrote their grammars.
Unfortunately ignorance of the past has equally spread to some scholars
writing about it. As a consequence, they may be led to draw wrong
conclusions. A case in point is an author who wondered why Tomâs Pinpin had begun
his 1610 Spanish grammar for Tagalogs with a chapter on numeral
expression. If he had known that the Spanish way of counting went against the grain
of the old Tagalog numerals, or vice versa, he would certainly have
elaborated better conclusions.
Indeed the old system is rather strange since many of its numerical values
are not expressed in a straightforward manner.

3.2. Basic principles

To begin with, if the multiplier of a power of ten is 1, pa^'rig is not used;
instead labfr sa is added before the multiplicand, then the product on its left
appears immediately after without the ligature (6)

(14) labfh sa raân isâh cf. jS^'ndaâ^'t isâh

one hundred and one (101) one L hundred and one
one L hundred and one

(15) labfh sa raân jsa^ngpuôi cf. jSa^^ j^p

one hundred and ten (11*) one L hundred and ten

(16) labih raân labîh'ng isâh cf.isâh'ndaâta't jSâj/mpû

one hundred and eleven (111) one L hundred and eleven
(17) labïh sa liboh dalawâh cf. iSâh'nlîboh't dalawâh
more than thousand two one L thousand and two
one thousand and two (1,002)
(18) labîh sa laksâi âpat na lîboh cf. jSâ^'nlaksâ 't âpat na liboh
more than 104 four L thousand one L 104 and four L 103
fourteen thousand (14,000)
The most surprising thing with the old system is that the multiplier is
incremented of one unit if it belongs to the set (2. ..8). To signal this change, the
prefix maika- (7) is added, in which ika- follows the same rules as the prefix
of ordinals. For thasake of clarity, I gloss maika- as « minus 1 ». The product
thus obtained is linked to the following one by the ligature.
(19) maika-tlôh'ng daân'ng siyâm
minus 1-three-L hundred-L nine
174

cf. dalawâjjiidaâj/t siyâm

two-L-hundred-and nine • •
two hundred and nine (209)

(20) maika-tlôh'ng liboh'ng maikalimâ^'ng daân

minus 1-three-L thousand-L minus 1-five-L hundred
cf. dalawâ^nlîbojj't âpat na raân
two-L-thousand-and four L hundred
two thousand four hundred (2,400)

(21) maikâpat na laksâq'ng labfh sa lîboh pitûh'ndaân

minus 1-four L 104-L more than thousand seven-L-hundred
cf. tatlôh 'ng laksâi at jSâj/nglïbo11 at pitûh'ndaân
three-L-10(4> and one-L-thousand and seven-L-hundred
thirty-one thousand seven hundred (31,700).
When the multiplier is 9, the expression moves one order upward, i.e. the
power (P) of M is incremented of one unit (Mp = Mp+1), and the multiplier
becomes 1: *maika- 9.MP = maika- l.Mp+1 = maika- Mp+1. Thus ten is
replaced by hundred, hundred by thousand and so on.
(22) labîh sa raân maikaraân'ng tatlôh
more than hundred minus 1-hundred-L three
cf. sân'ndaân't siyâm-na-pû 't tatlôh
one-L-hundred-and nine-L-ten and three'
one hundred and ninety three

(23) maikaliboh'ng wal6h

minus 1-thousand-L eight
cf. siyâm na raân't walôh
nine L hundred-and eight
nine hundred and eight.
Once the unit of the following order is reached, the system is similar to
the new one.
(24) dalawâj/ng daân two-L hundred two hundred.

3.3. Erasure of «ten»

The system is compounded by the fact that the word standing for the first
power of ten (puô* is erased.
In the series from 11 to 9, sa puô* (than ten) disappears and 'ng is added
after labft. Habfi sa puifi ml = labfiO 0 L m I. As intimated in the second
section, the new system borrowed this "formula from the old one.
(25) *labîh sa puô^ dalawâh - labfh'ng dalawâh
more than ten two = more than-L two
twelve.
When the multiplier of the power of ten belongs to the set (2.. .8),
disappears: maika- m + 1 10 L ml = maika- m + 10 L ml.
(26) maika-tlôh'ng dalawâh minus 1-three-L two
cf. dalawâhmpûq't dalawâh two-L-ten-and two
 175

twenty two.
Thus from 20 to 29, some of the expressions are:

(27)
20 dalawâ^mpuôi two-L
21 maikatlôh'ng isâh minus 1-three-L one
25 maikatl6h'ng limâh minus 1-three-L five
29 maikatlôh'ng na siyâm minus 1-three I nine

When the multiplier of puô* becomes 9, the power of pué* being

increased of one unit, the expression switches to hundred: *maika- 9.10(1) L ml =
maika- 1.101+1 = maika- 1.10<2>.
(28) maikaraan'ng isâh minus 1-hundred-L one ninety one.

34- An algorithm
What I have said so far does not fully explain all the numeral expressions
in the old Tagalog system for, comparing the two following examples.
(29) maikalibôh'ng siyâm na puô^
minus 1-thousand-L nine L ten
cf. siyâm na raân't siyâm na pû^
nine L hundred-and nine L ten
nine hundred and ten

(30) maikalïboh'ng maikaraân'ng siyâm

minus 1-thousand-L minus-1-hundred-L nine
cf. siyâm na râan't siyâm na pûq't siyâm
nine L hundred-and nine L ten-and nine
ninei hundred and ninety-nine.
shows that for a given product the shift to the order above is triggered by
the product immediately on its right.
I therefore propose the following algorithm. I posit that the numerical
expression starts from a node Y. This node branches into two sub-nodes X
and Y. Then every sub-Y branches again into X and Y until the rightmost
product is ml, as defined in section (1). Thus 96,185 will be analysed as
9.10* + 6.103 + 1.102 + 8.10 + 5.1
176

and expressed as
(31) maikayuta 'ng maikapitôh'ngliboh'ng labîh sa raân'ng maikasiyâam na limâh
minus 1-hundred thousand-L / minus 1-seven L-thousand-L / more than hundred-
L I minus 1-nine / L /five
ninety-six thousand one hundred and eighty-five (96,185).
Tagalog numeral expressions vary according to whether Y is nil or not,
and whether the M of X (Mx) is equal to 1, belongs to the set (2...8) or is equal
to 9.

3.5. Truncated expressions

Another feature of the old system is that any power of ten may be deleted
if the following Y is nil. Examples are to be found in San Joseph (1610 [1752]:
666,669) and Laktaw (1929:fn. 2,3,4). This truncation generates ambiguity.
Of course any product thus deleted pied-pipes its ligature.
(32) labfh sa raân isâh
one hundred and one (101).

(33) labîh sa raân isâh = labîh sa raân i

one hundred and ten (110)

(34) maikatlôh'ng lîboh'ng tatlôh = maikatlôh'ng lîboh'ng maikapât na raân

two thousand three hundred (2,300)
(35) labih sa laksâ** isâh = labîh sa laksâ^ isâ'ng lîboh
eleven thousand (11,000).
If the name of the object counted is involved, the numeral expression
cannot be truncated.
(36) labîh sa raân isâj/ng kawâyan
one hundred and one bamboos (101)

(37) labîh sa raân isâh'ngpuôq'ng kawâyan

one hundred and ten bamboos (110).

3.6. Use of the connector

Strangely enough, the connector at appears in the examples given by
Laktaw.
(38) labîh sa lîboh isâ h at isâ^ng kawâyan
labîh sa lîboh jSa^ngdaân at isâh'ng kawâyan
one thousand one hundred and one bamboos (1,101)

(39) labih sa lfboh isâh at labfh'ng isâh'ng patakyâ^

labîh sa lfboh jSa^'ngdaân at labfh'ng isâh'ng patakyâ^
one thousand one hundred and eleven betel boxes (1,111).
At already appears in some of San Joseph's examples.
(40) labi sa libon ysa, at sang cauayan (p. 668)
 177

labih sa liboh'ng isâh at isâ^ng kawâyan

one thousand one hundred and one bamboos (1,101)

(41) dalauang daan, at ysang cauayan (p. 668)

dalawâjj'ng daân at isâj/ng kawâyan
two hundred and one bamboos (201).

Parameters Numeral Example

Numerical
P Y mx

I ml m 5: limâh

1 0 m.10 m L puôq 50:

limâh'ngpuôq

> 1 m.10 + labîb sa Y 15: labih sa

limâh
0 Y

2...8 m.10 + maika- 45:

Y m+1.10LY limâh
maikalimâh'ng

9 m.10 + maika- m.lO1+1 95:

Y LY
isa^nda^ng
limâh

m.lOp
> 0 mLM 500:
limâh'ngdaân
1
> 1 m.Kf labih sa 10* Y 105: labïh sa
0 + Y raân isâh

m.lOp maika-
2...8 405:
+ Y m+l.Kf + Y maikalimâh'ngd
limâh
aân'ng
nUO* maika- l.KT1
9 905:
+ Y + Y
limâh
maikalïboh'ng
178

So, differently from what we have in the modern system, at plays a minor
role in the old one. It became ubiquitous from the moment the prefix maika-
became obsolete.

3.7. Miscellanies
Apart from cardinal, ordinal and fractional numerals, San Joseph
provides a fairly comprehensive study of various other numeral forms such as
restrictives, e.g. tatatlô h / tatlô h -tatlo h « only three », repartitives, e.g. dalâ-
dalawâ h « by two », distributives, e,g. tiglilimâ h « five each »,
multiplicatives, e.g. makalimâ h « five times », minsan « once », proportionates, e.g.
makapupuô i « one (part of honey) for ten (parts of rice) ».
In addition several pages are devoted to currency and measure.
Conclusion
One might wonder why pre-hispanic Tagalogs had developed such an
intricate system. It looks so awkward that one is reluctant to admit that it was
actually used in daily transactions. Yet, it must have been since it is
described as such in early grammars.
This is all the more puzzling as pre-hispanic Filipinos did not have digits.
The explanation might lie in the observation made by San Antonio (1738 [Picor-
nell's translation 1977 : 165] that Tagalogs computed with little pebbles, and
recorded the result in words written in their baybayin syllabary for they had
no scripts for digits. The friar might have specified that, differently from
Arabic, Greek and Hebrew letters, Philippine syllabic signs had no numerical
values.
My interpretation is that computation and numeral expressions were
entirely separated. The former depended on an abacus (8) drawn with a stick on
the ground which, for all its crudeness, was in no way inferior to any figure
drawn on a blackboard by a mathematics teacher. The use of small pebbles
as tokens should not deter us from concluding that this technique must have
been fairly sophisticated; after all, didn't calculus (reckoning) mean « small
pebbles » in Latin ? Once the result was obtained in this silent (?) way, it was
read aloud and/or taken down in their syllabic script.
The Tagalog abacus must have been based on a system whereby
quantities were to be reached rather than completed. For example, 47 was
perceived as 50 minus 3, hence the presence of 5 tokens in the column for tens.
I also surmise that the original abacus had been devised for the nonary system
- in which nine units of a given order make one unit of the order above -
hence the necessity to pass to the following column when the figure was 9,
even though the Tagalogs reckoned in the decimal system. Of course it would
remain to account for the use of a nonary abacus with the decimal system
in spite of the discrepancy involved. Until we find new documents, however,
it would be idle to go further into this speculation. Therefore I shall leave
the matter as it is - and to the professional mathematician.
Another striking feature is such large orders as 10 ro . They come as a
surprise in so far as pre-hispanic Tagalogs were deemed to live a rather
primitive life. One cannot imagine that such high powers of ten were of any use
in their daily activities. Yet there must have been people who needed them
since these were recorded in dictionaries.
 179

We must therefore infer that these were necessary only in speculative

calculations involving very large numbers with no practical applications. One
cannot help thinking of a deep Hindu influence (through Java) where such
numbers mainly served in the calculation of the years to the spent in the
netherworld by unfortunate sould (see Feer 1892). If this view could be
supported by other proofs, we might have some day to reform our vision of pre-
hispanic Philippines. If we align these stupendous figures and religious words
such as sampalatayah (« faith », seeDoctrina Christiana, 1593) borrowed from
Sanskrit samrpratyaya (Monier-Williams : 1174), we may conclude that during
the pre-Hispanic period there were Filipinos whose religious and
arithmetical views were so sophisticated as to lead them to imagine extremely large
quantities and extremely long periods beyond any man's experience. All these
ideas shed an odd light on the lost centuries of Philippine history.
Now, if the Spaniards had such a poor opinion of Filipinos, some
important event must have taken place some time long enough before their arrival
to account for the sd state in which they found them. Because of the recent
Pinatubo eruption, one surmises a natural catastrophe that would have swept
a budding civilisation, though large-scale looting, massacres and slavery by
pirates from the area now known as Indonesia should not be discarded. Only
archeology could solve this problem.

NOTES
1. The ligature is a linker that connects two items belonging to the same syntactic group. It
is realized as -ng after [n], [h] and [?], to which it is substituted, and na in all other cases.
2. So far, no document has ever yielded any word for « hundred million ».
San Joseph (1610 [1752] : 669) does not know any numeral beyond yûtai (105). He then
enters the word bahdlcfl in the following manner. «Millares de yota, no se conoce: sino dizen
sang bahala, que es dezir vn que se yo, ycao na an bahala, echa por essos trigos de Dios
que ya no se puede pensar». (op.cit. 669-670): «[the expression for] «thousand yûtaq» is
unknown, instead they will say «pâ^'ng bahâla'1» [one bahdkfl], which somehow means «Ikâw
nd ang bahdkfl», taking the wrong road, which is unthinkable».
The passage is all the more opaque as Ikâw nâ ang bahâkfl currently means «You decide.
/ You take care of everything». Could bahakfi, be a misprint? Could the right word be balaga,
which San Joseph uses p.675 in the expression sasangbalaga (only one balaga)"! This is not
impossible for the latter cannot be confused with halagâ)1 «value» in so far as a comparison
of its initial with the B of labi in the same page stultifies this qualification. Unfortunately,
the only balaga entered in existing dictionaries is balagâ^ «boo» (Panganiban 1972). Thus
the numeral used for 108 is still a mystery.
3. Gates is nos mentioned by San Joseph (1610), but is entered in Laktaw's dictionary (1914).
4. The anteposer makes it possible to move the focused item to the beginning of the clause,
e.g. Lumdlangoy si Pedro. I Si Pedro ay lumàlangôy. « Pedro is swimming » This is only
a stylistic change, characteristic of the written language.
5. Joseph [xo'se] is now spelt José.
6. San Joseph's sometimes suggests the presence of the ligature, e.g. libon instead of libo in
labi sa libon ysa I labih sa libov/ng isâh «one thousand and one» (p. 667), lacsan instead of
lacsa in labi sa lacsan ysa / labih sa laksâVng isâh «eleven thousand» (p. 669).
7. Mai is pronounced [mei]. San Joseph spells it now may -obviously conservative - now rney
-probably a representation of what he heard -, and Totanes consistently uses mey, which
shows that the prononciation has not varied for the last four centuries.
180

A simple abacus is a board divided up into one hundred squares on which tokens are placed
according to the operation performed. The portable abacus substituted beads on rods. Some
Roman abaci were pocket models with cursors sliding in grooves. Nomographs allow
complex calculations such as volumes. Pythagoras (ca. 580-496 BC) is said to be the inventor
of the abacus. The Chancellor fo the Exchequer was thus named in Anglo-Norman because
his abacus was reminiscent of a chessboard. The board abacus was used in France until the
beginning of the 19th century. For instance in the opening scene after the ballet, the lead
character Argan in Molière's Le malade imaginaire [The hypochondriac] (1673) uses this
instrument to make his reckonings. Many stage directors do not seem to be aware of this.

BIBLIOGRAPHY
Anonymous (1593). Docrina Christiana (en lengua espaiïola y tagala corregida por los Religio-
sos de las ordenes [A catechism]. Manila (Impressa con licencia, en S. Gabriel de la orden
de S. Domingo), 76 unnumbered pages.
Anonymous (s.d. [1926]). Casaysayan nang Abecedario [A history of the alphabet]. Maynila,
Martinez. 55 + 7 pp.
Anonymous (1980). Petite Encyclopédie des Mathématiques[A small encyclopedia of
mathematics]. A French translation of Kleine Enzyklopàdie der Mathematik, Leipzig, 1975, 828 pp.
Bloomfield, Leonard, 1917 [Johnson Reprint: 1967]. Tagalog texts with grammatical analysis.
Urbana, Illinois, University of Illinois Press, 408 pp.
Bouzet, Jean (1978), Grammaire espagnole [A Spanish grammar]. Paris, Belin, 434 pp.
Coedes, G,. (1930). « Les inscriptions malaises de Ç rivijaya » [The Malay inscriptions of Shrivi-
jaya] in Bulletin de l'Ecole Française d'Extrême Orient (Paris), 1930, pp. 29-80.
Dell, François (1981). « On certain sentential complements in Tagalog », Cahiers de
Linguistique - Asie Orientale (Paris), 10, pp. 19-41.
Dempwolff, Otto (1934-1938). Vergleichene Lautlehre des austronesischen Wortschatzes. Berlin,
Beiheft zur Zeitschrift fur Eingeborenen-Sprachen. 3 vol. English translation (1971) :
Comparative phonology ofAustronesian Word Lists. Quezon City, Philippines, Ateneo de Manila
University. Vol. 1 : 134 pp.; Vol. 2 . 342 pp.; Vol. 3 : 224 pp.
Feer, Léon (1892). « L'enfer indien » [The Indian hell] in Journal Asiatique (Paris) 1892, pp.
185-232.
Ifrah, Georges (1981). Histoire universelle des chiffres [An universal history of digits and
figures]. Paris, Seghers, 568 pp.
Karush, William (1989). Dictionary of mathematics. Webster's New World, New York, 317 pp.
Laktaw, Pedro Serrano (1914 [1965]. Diccionario Tagdfog-Hispano [Tagalog-Spanish dictionary].
Manila, Santo y Bernal-Echagûe. Reimpression : Madrid, Ediciones Cultura Hispanica. 2
vol., 1392 pp.
Laktaw, Pedro Serrano (1929 [posthumous]. Estudios gramaticales sobre la lengua tagalog
[grammatical studies on the Tagalog language]. Manila, Fajardo, 370 pp.
Llamzon, Teodoro, 1%76. Modem Tagalog : ajunctùmal-structural description. The Hague :
Mouton. 131 pp.
Marre, Aristide (1902). Grammaire tagalog composée sur un nouveau plan. Leyden, the
Netherlands, overdrukt uit [offprint from] de Bijdragen tot de Taal-, Land- en Volkenkunde van
Ned.-Indië, 6e VolGr. IX, 47 pp.
Monier- Williams, Monier (1899 [1981]. Sanskrit-English dictionary. London, Clarendon Press,
1.333 pp.
Morga, Antonio de (1609). Sucesos de las Mas Filipinos [Events in the Philippine Islands]. Rizal
edition, Paris 1890. Centenary offset impression, Manila, 1961, 374 pp. Translation into
Tagalog by Pambansang Komisyon ng mga Bayani [National Commission of Heroes], Maynila
(Manila), 1964, 371 pp.
Panganiban, José V. (1972). Diksyunaryo-Tesauro Pilipino-Ingles. Quezon, Philippines : Manla-
paz. 1027 pp.
 181

Pinpin, Tomâs (1610). Librong pagaaralan nang manga Tagalog nang uicang Castilla [A book
for the study of the Spanish language by Tagalogs], Bataan, Philippines : Diego Talaghay.
Potet, Jean-Paul (1987). « La pétition tagale Caming manga alipin (1665) ». [The Tagalog
petition1,Caming
n° juin 1987,
manga
pp. alipin
109-157.
(1665) in Cahiers de Linguistique-Asie Orientale, Paris, Vol. XVI,
San Antonio, Juan Francisco de (1738). Chronicas de la Apostolica Provincia de S. Gregorio de
Religiosos Descalzos de N.S.P.S. Francisco en las Islas Filipinos, China, Japon, &c.
[Chronicles of the apostolical province of Saint Gregory...] Translation : Picornell, Pedro (1977).
The Philippine chronicles ofFray San Antonio. Historical Conservation Society, Manila, 265
pp.
San Joseph/José [modern spelling], Francisco de (1610 [1752]). Arte y reglas de la lengua tagala
[A grammar and rules of the Tagalog language]. Bataan. xvi + 311 pp.
Schachter, Paul & Otanes, Fe. 1972. Tagalog reference grammar. Berkeley : University of
California Press. 566 pp.
Totanes, Sebastian de (1745 [1796]). Arte de la lengua tagala [A grammar of the Tagalog
language]. Sampaloc, xiv + 247 + 3 pp.