Establishing genetic relationship - comparative method

Question 1

Languages may be similar as a result of

Answer 1

1. nature
2. chance
3. borrowing/diffusion (i.e. language contact)
4. inheritance from a common ancestor

MUST RULE OUT ALL OTHER POSSIBILITIES to be certain langs are from COMMON ANCESTOR

Question 2

Common areas of Natural Resemblance

Answer 2

ONOMATOPOEIA
cross-ling tendency for wordforms linked to certain kinds of meanings to approximate with real world sounds with which they are associated. - degree of ICONICITY

NUSERY FORMS
‘Mama/Papa’ words also share a high degree of cross-linguistic similarity regardless of common ancestry - e.g. often INVOLVE LABIALS.

Question 3

Natural Resemblance in regards to Structural Similarity

Answer 3

Paradigms, syntactic frames etc. = DIAGRAMS and as such they are largely iconic.

This means that similarity in the structure of paradigms or syntactic frames may be natural and completely independent of genetic relationship.

Question 4

Factoring out natural resemblance

Answer 4

Natural resemblance is ruled out if comparison is restricted to symbols:

signs in which the relation between signans (signifier) and signatum (signified) is arbitrary and conventional
(simplex) lexical signs (i.e. words)
HG Hund ‘dog’
NE hound ‘k. of dog’

Question 5

Chance resemblances

Answer 5

Conventional wisdom holds that 5-6 per cent of the vocabulary of any two compared languages may be accidentally similar.
So, if only 5-6 per cent of the overall vocabulary compared is similar, we cannot conclude anything …

STRUCTURAL SIMILARITY - poverty of choice
There is a 50/50 chance that two languages will have the same order, whether or not they share common history (if a language has an adjectival phrase)

Question 6

Factoring out chance and borrowing

Answer 6

Chance and borrowing can be ruled out if the similar word forms display regular sound correspondences.

• Sound correspondences in chance resemblances are not regular - - except when there has been heavy borrowing from a single source (heavy borrowing situations are therefore problematic).

Question 7

Regular Sound Correspondences

Answer 7

L1 and L2 show regular sound correspondences iff, whenever p[i] occurs in a word in L1, p[j]occurs in the corresponding position in a related word in L2, where ‘related’ implies ‘similar in meaning’…

REGULARITY ASSUMPTION is a FUNDAMENTAL PREMISE of the COMPARATIVE METHOD

Question 8

Summary of Establishing Genetic Relationship

Answer 8

Genetic Relo = HIGHLY LIKELY when two/more langs display REGULAR SOUND CORRESPONDENCES in LEXICAL ITEMS that are similar in Both FORM and MEANING

• demonstrate these lexical items = likely RETENTIONS from proto-lang that diversified into seperate languages
• Lexical Items w reg sound correspondences = COGNATES
• identify via COMPARATIVE METHOD

Question 9

What does the comparative method rely on?

Answer 9

The comparative method relies on the identification of sets of POSSIBLE COGNATES (lexical look-alikes) and proceeds by determining whether or not these are REAL COGNATES.

The lexical look-alikes should be similar in both:
form (i.e. phonologically)
meaning (i.e. semantically)

Question 10

How similar do cognates have to be?

Answer 10

Potential cognates dont have to be similar in LITERAL sense, they have to be Similar given our knowledge of what may be RELATABLE SOUNDS and RELATABLE MEANINGS

EG. /p/ and /w/ are similar because we know that /p/ is a possible source for /w/ via a process of lenition.

EG. to want’ and ‘future tense’ are similar meanings because we know that ‘want’ is a possible source of ‘future tense’ via a grammaticalization process.

Question 11

Comparative Method Process

Answer 11

1. identify potential cognates
2. confirm with Regular Sound Correspondences
3. extract CORRESPONDENCE SETS from data to see if recurrent
4. check distribution of correspondence sets that involve the same/v similar segments - find environment for conditioned change
5. each set of complementary regular correspondences reflects a phoneme in proto-language - RECONSTRUCT

Question 12

Correspondence set definition

Answer 12

a correspondence set is a set of phonological units occurring in corresponding positions in that data set, e.g

Question 13

Two forms are cognate iff:

Answer 13

1. Their meanings can be related
2. EACH and EVERY component of their forms (allowing for morph differences) is in REGULAR correspondence

• some correspondences might involve Ø

Question 14

Assigning a phonetic value to a proto-phoneme

Answer 14

We choose a phonetic value which presupposes most plausible sound changes in the daughter languages.

Much less important …
Occam’s Razor: given equal level of plausibility, we prefer the reconstruction that requires the fewest acts of change.