Languages In the World (30 page)

Philologists need to know where to look in the vocabularies and structures of the languages in order to begin the reconstructive process. They want highly stable words and constructions presumably inherited from an imagined protolanguage, not words and constructions borrowed through contact with speakers of languages from a different family. Likely categories of stable vocabulary include the names for: family members and relationships, body parts, features of the natural world (sun, moon, sky, etc.), numbers one through 10 at least, and common animals and plants indigenous to the homeland.

Likely categories of stable structural features include:

1. The sound system and syllable structures: in Chapter 5, we noted that the Greeks needed to turn some of the Phoenician symbols into vowels because Indo-European languages characteristically have syllables with a large number of consonant clusters. They also have midsized phonemic inventories. By way of contrast, Austronesian languages avoid consonant clusters and have relatively small phonemic inventories.
   
2. The pronoun system and what kinds of distinctions are made: singular, dual, plural, etc., as well as the words for the pronouns themselves. Hawaiian, for instance, has we-inclusive
   kakou
   ‘all of us including you' and we-exclusive
   makou
   ‘all of us excluding you,' and this distinction seems to have been a feature of Proto-Austronesian, since it is found throughout the Austronesian-speaking world from Hawaiian in the east to Malagasy in the west and important points in between, including Indonesian and Tagalog.
   
3. Word-formation preferences: Swahili favors agglutinative morphology by prefixing, while Turkish favors agglutinate morphology by suffixing, as mentioned in Chapter 3. A widespread characteristic of Austronesian languages is their use of infixing, which involves inserting a personal pronoun morpheme and/or a tense morpheme into the word itself. Tagalog has borrowed the English word
   graduate
   . To say the past tense ‘I graduated,' the infix
   -um-
   is inserted near the beginning of the word
   gr
   um
   aduate
   .
   
4. Word-order patterns: many, but not all, Austronesian languages have verb-initial word order. Language stocks and families tend to have preferred word orders, but one of the many structural features that can change over time in a language is dominant word order. Nevertheless, widespread preferences for word order in a family and stock are common. The Dravidian languages, for instance, all tend to be SOV. It is not surprising that Austronesian languages all show a preferred VSO order.
   

Instabilities in some of these more-or-less stable categories can always be found. Although pronouns are not likely candidates for replacement, speakers of English gave up their West Germanic third person pronouns along with the plural forms for the verb ‘to be,' and they borrowed
they
,
them
,
their,
and
are
from Old Norse. Numbers can be revealing. When a language has two ways to refer to numbers, this is often a sign of long domination by another language. The English ordinal
second
is not based on the English cardinal
two
but is borrowed from French. The Vietnamese numbers begin:
mọt
,
hai
,
ba
,
bốn
 … , and the days of the week begin:
thứ hai
meaning ‘day two,' namely Monday, and
thứ ba
meaning ‘day three,' namely Tuesday. However, the word for Wednesday is
thứ tư
meaning ‘day four' (instead of
thư bốn
), after which the names for the days of the week revert to compounds with the Vietnamese cardinal numbers. Additionally, in Vietnamese the word for an intersection of two streets is
ngã tư
, meaning ‘crossroads/corners four.' The word
tÆ°
‘four' is a Chinese borrowing.

Apart from these minor observations, we can say that the
borrowing
patterns we see today are probably those that have existed for all time, and when we eliminate those, we get down to the stable vocabulary. In English, we easily spot borrowings in categories such as food (
sushi
,
taco
,
wurst
), clothing (
kimono
,
serape
,
dirndl
) and specialized domains such as opera, which gives us Italian words such as
soprano
,
alto
,
vibrato
, and so forth, along with the word
opera
itself.

The Old English plural of ‘cow' was
kine
. The Modern English plural is
cows
. No plausible regular sound change can be invented to account for this change. Only from the reformation of the noun classes during the Middle English period, when the -s# plural that was on 35% of Old English nouns began to generalize, can the change from
kine
to
cows
be explained. The plural
cow
s was reformed by analogy to the many other nouns adopting the -
s
plural at that time. Analogy can also be called
pressure of the system
. When children learning English say “she bringed me a cookie,” they are forming the past tense by analogy to the many other past tenses they know that form a system. Eventually they adjust to the normative and irregular form
brought
.

The example of
bringed–brought
demonstrates how the principle of shared irregularity as a mark of common descent and the workings of analogy lock into one another. In Modern German, ‘to bring' is
bringen
, and the past tense ‘I brought' is
ich brachte
. In Modern German, ‘to think' is
denken
, and the past tense ‘I thought' is
ich dachte
. Despite the fact that English and German have verbs like
ring–rang–rung
/
ringen–rang–gerungen
and
drink–drank–drunk/trinken–trank–getrunken
, the
verbs ‘to bring' and ‘to think' in both languages share closely related irregular past tenses. They must have inherited these irregularities rather than independently inventing them in just those words. This shared irregularity is thus (one of) the mark(s) of the common descent of English and German as West Germanic languages. At the same time, the philologist has to be on the lookout for forms that speakers have analogized, that is, regularized over time, and be willing to invoke the working of analogy to override the principle of the regularity of sound change. The principle of shared irregularity marking common descent and the working of analogy help the philologist understand the contours of the jigsaw puzzle pieces used to reconstruct the protolanguage.

The reconstructed PIE word for ‘daughter-in-law'
^*
snusós
makes an interesting example.
¹⁰
The
-ós
is of interest because it is a masculine ending reconstructed for a clearly feminine entity. The reason for the choice of the ending comes from the forms for the attested words in Latin
nurus
and Greek
nuos
, which both have masculine endings. All other Indo-European cognates have a feminine ending. So, Greek and Latin must have inherited the shared irregularity of a feminine entity carrying a masculine ending. All other languages must have analogized the ending to make it conform to other feminine nouns, because there is no sound change that will account for why the other languages changed an -os# to an -a#. This reasoning makes more sense than imagining that Greek and Latin inherited an originally feminine ending and both independently and for no apparent reason decided to give it a masculine ending.

Long before there were language academies and language police, societies as a whole placed restrictions on words. Today, the words for certain body parts and body functions tend to come under taboo, as well as do so-called four-letter words. No sound laws will account for the phonetic shape of the nearly endless string of euphemisms in English for words under taboo:
dang
,
darn
,
durnit
,
fooey
,
fudge
,
gee
,
gee whiz
,
gee willikers
,
golly
,
gosh
,
heck
,
jeez
,
sam hill
,
shoot
,
shucks
, … the list is long. The only way to account for how they sound is by how well they deform the word under taboo, thereby diminishing its strength, while still keeping some part of it audible.

The names for things to be feared and respected also need to be altered, so that one does not draw the power and the wrath of the thing named. (Don't take the Lord's name in vain!) Medical researchers may refer to cancer as The Big C. The M-word is uttered at a confirmed bachelor's peril. The L-word has a variety of interpretations. Sometimes, whole words are replaced. In modern Judaism, the word
Adonai
refers to the Lord because the name
Yahweh
is not to be pronounced. Some times, the words are not uttered in the first place. “I'm hoping for a good outcome for my grant application, but I don't want to tell you to which funding agency I submitted it, so as not to jinx it.” We don't really and truly believe in the magical power of words (do we?), but sometimes we behave as if we do.

The Germanic, Baltic, and Slavic branches of Indo-European presented philologists with a puzzle concerning the question of whether or not the Proto-Indo-Europeans knew about bears. It seemed that they lived in an area that had bears, and enough IE branches had words for bears that made the reconstruction
^*
arkth- possible for ‘bear.' However, the northern tribes – the ones presumably encountering bears fairly
regularly – had words for this animal such that no amount of sound correspondences could lead back to
^*
arkth
-. The problem was resolved by appealing to the notion of taboo replacement. Each language found a way to refer to bears without invoking them, and they did this by nicknames. The Germanic peoples called them some form of ‘bruins,' that is, a ‘brown one,' making
bear
and
brown
cognates. The Baltic people referred to them as a ‘shaggy one,'
lācis
, while the Russians called them a ‘honey-eater,'
medved
, which is also a Russian family name. The Bear itself is the symbol of Russia.

We have spoken of population genetics and linguistic reconstruction, but thus far we have conjured without the help of cultural anthropology or archeology. Part of the interest of philology is that reconstructed languages give glimpses into the cultural and intellectual conditions of the speakers in a way that the archeological evidence provided by burial mounds or pottery shards cannot. What words can and cannot be reconstructed for a protolanguage reveal much about the local conditions in which the speakers of that language lived, the material culture they possessed, and even some of the abstract and/or religious and political ideas they held. At the same time, cultural information can and at times must be imported into the reconstructive enterprise. Albanian
nu
means ‘bride.' It was used to help reconstruct PIE
^*
snusós
, despite its phonetic divergence from the original form, because there is ample evidence that Indo-European culture was highly patriarchal, and a bride would leave her family home and move in with the husband's family, thus becoming at the same time a daughter-in-law. Thus,
nu
is deemed an appropriate cognate. For certain Native American reconstructions, the words for ‘hair' and ‘feather' might be legitimate cognates.

Philologists bring long-lost human worlds into existence by organizing available facts by means of tested principles and feats of logic. Historical and comparative philology is a sometimes frustrating, often challenging, and always fascinating exercise in ingenious inference.

Some of this ingenuity is mobilized for the purposes of determining the historical homelands of the speakers of the various protolanguages. Archeologists have long worked in tandem with cultural anthropologists and linguists, and more recently with population geneticists to determine the most likely locations for these homelands. It is generally agreed that Formosa/Taiwan was the incubator for Proto-Austronesian. However, in the case of the homeland of the pre-Proto-Austronesians, the possibility of their settlement around Hangzhou Bay before arriving in Formosa/Taiwan is cast in some doubt by the cultural fact that in this area of China, the agricultural staple is rice, while the Formosan's is millet, and the Formosan culture has an agricultural rite the heart of which is millet. On the other hand, a word for ‘rice' can be reconstructed for Proto-Austronesian. Lining up what can be imagined for the past with what is known in the present is a tricky business.