when empirical evidence meets linguistic principle evidence should win every time!

18
WHEN EMPIRICAL EVIDENCE MEETS LINGUISTIC PRINCIPLE Evidence should win every time!

Upload: hugh-holt

Post on 17-Dec-2015

217 views

Category:

Documents


0 download

TRANSCRIPT

WHEN EMPIRICAL EVIDENCE MEETS LINGUISTIC PRINCIPLE

Evidence should win every time!

Evidence vs. Linguistic Principles

Evidence trumps principles in science. But the relationship between the two is interesting.

Principles are used to decide between alternative explanations.

For example, simplicity comes into play when one solution can be said to be simpler than another solution. Other than that, simplicity has no meaning whatsoever.

On Dogmatism

Dogmatic is what we call a person who sticks to principle despite the evidence.

“That solution is too complicated” is meaningless unless there is, in fact, a simpler alternative.

Consider the structure of the human eye. In what sense can it be called “too complicated”?

Two uses of those principles

In addition to using principles to choose among alternatives that are presented to us, there is another purpose to which they can be put, namely they become part of a discovery procedure.

Principles and Discovery in Science

When “something tells us” that an evidence-based solution is suspicious because it is “too complicated” or “violates phonetic plausibility” or some other principle, that “something” rightly causes us to search for alternatives.

But if an alternative is not found, it is sheer dogmatism to abandon the evidence-based (i.e. viable) working hypothesis merely on a vague feeling that there ought to be a better solution.

Example: Lenition vs. Fortition

All things being equal, lenition is the more expected type of sound change on grounds of phonetic plausibility, which is a form of simplicity argument.

However, the Second Consonant Shift in modern German was fortition:

p > pf, t > ts, k > x.

The shift accounts for English penny vs. Pfennig, to vs. zu [tsu], and make [-k] vs. machen [-x-].

When fortition makes sense

The High German shift occured syllable-initially, and not after a vowel. This is the same position where stops are aspirated in English.

tin [th], often [th], attend [th] vs. atlas [t], atmosphere [t].

Q.E.D.

If we had nothing other than English and German to go on, we might well conclude that *pf > p in English. But we know on the basis of other languages that it was actually the other way around: *p > pf syllable-initially in modern German.

In other words, fortition happens.

Another principle: Uniformitarianism

Uniformitarianism is a form of argument that favors the use of familiar phones (as defined by the language type) vs. exotic ones to achieve the same purpose.

But what if all the alternatives violate Uniformitarianism?

Examples can be found in almost every language family.

So much the worse for Uniformitarianism.

Example: PAn/PMP *j

PAn/PMP *j is hypothesized to have been a palatalized voiced velar stop *[gy], which no daughter languages exhibit directly.

The reflexes are either /k,g/ or /t,d/ in language after language; moreover, reflexes are only found word-finally and in intervocalic position.

Melanau /t,d/=Rejang /k,g/

PM Rejang-MusiPMP *quləj *ulət olok

“caterpillar”

PMP *qapəju *əmpədo pəgəw“gall bladder”

Does that mean we throw Uniformitarianism out the window?

WHA--?

NOT AT ALL

WE WORK WITH IT!

PMP *j is a mystery

This means employment for the coming generation!

Blust’s challenge to Neogrammarian principles

Noting perhaps that the greatest names in linguistics (Saussure, Verner, Chomsky) got that way by breaking the mold, Blust’s work is full of surprises.

Here is a recent article worth reading both for its accurate description of Historical-Comparative linguistic principles and for its use of evidence to challenge to those same principles.

 C = Conditioned; U = Unconditioned; PA = Phonological alternation; NG – New Guinea; Jav = Java; Br = Borneo; Oc = Oceanic; S = Sumatra; Sul=Sulawesi (Celebes)  1. *w/y > -p in languages of western Manus (NG) C no PA 2. *w/b > c-, -nc- in Sundanese (Jav) C no PA 3. intervocalic devoicing in Kiput (Bor), Berawan (Bor) C no PA 4. *dr>kh in Drehet (Oc) U? no PA 5. *b/d/g>-m/n/ŋ in Karo Batak (S), Berawan (Bor) C no PA 6. C>C:/__V# in Berawan (Bor) C no PA 7. *b> -k- in Berawan (Bor) C no PA 8. *g> p-, -j-, -p in Sa’ban (Bor) C no PA 9. *an/aŋ>-ay and *em/en/eŋ> -aw in Iban (Bor) C no PA10. postnasal devoicing in Murik (Bor), Buginese (Sul) C no PA 

Figure 1. Some bizarre sound changes in Austronesian languages

Note: Berawan is closely related to Melanau according to Ray (1913).

Must sound change be linguistically motivated? (Blust 2005:221)

LING 484/585WINTER 2009

WHEN EVIDENCE MEETS PRINCIPLE IN HL