Comparative Methodology

[Steve Long]

----------------------------Original message----------------------------
In a message dated 4/30/02 10:31:34 PM, lvhayes at worldnet.att.net writes:
>As I understand things at this point, the process appears to be divisible
>into at least the following phases:
>1) Data collection.
>2) Data analysis.
>3) Application of the comparative method.
>4) Reconstruction
>5) Presentation.

In a message dated 5/1/02 9:43:07 PM, lvhayes at worldnet.att.net writes:
>Please note the distinction I make between 'phase' and 'step'.  A phase is
>a stage in the implementation during which one or more (related) steps are
>performed, a step is a specific process conducted within a phase.  I also
>didn't include a phase that would cover situations where the languages
>compared split off the ancestral node at different times, such that
>reconstruction of one or more intermediary proto-languages might be
>necessary.  This phase would also require iteration of at least some of
>the preceding phases.

I think that, whether you are calling it steps or phases, it may be valuable 
to consider that the raw collection of data could be directionless unless 
some guiding hypothetical (confirmed or unconfirmed) is there from the start 
of the process.  Obviously, a comparativist is going to be looking for data 
that shows systematic correspondence in comparing two languages (or its 
absence) and that theoretical structure should affect what is gathered.  The 
mass of data available in the computer age accents just how much 
pre-selection must go on in any scientific endeavor and makes data screening 
methodology an important issue in any field.

What has helped me understand these early issues a little better is the 
literature on field linguistics.  If you can get a hold of some of these 
kinds of publications, the more "deductive" parts of the early part of the 
process I think become clearer.

>From that perspective, the following was helpful on basics:
Samarin, William J. 1967. Field Linguistics: A Guide to Linguistic Field Work.
New York: Holt, Rinehart and Winston.

There are harder to get monographs from the Linguistic Society that generally 
address the specific issue (these two were at the main branch of the NYPL):

1967 "Lahu and Proto-Tibeto-Burman:the utility of non-written languages for 
comparative reconstruction." Linguistic Society of America, Ann Arbor, 
Michigan, July 28.

1990 "The relationship between empirical and theoretical linguistics." [De 
relatie tussen empirische en theoretische taalwetenschap] Symposium to 
celebrate the opening of the Institute for Descriptive and Comparative 
Linguistics, University of Leiden (Netherlands), November 7.

The hypothesis that languages are related (or unrelated) as a starting point 
is a consciously addressed issue in some field work that deals with a mass of 
data that may have very old connections, especially where areal and contact 
issues are also present.  The following book really accents this, if you can 
get a hold of it:
Abbi, Anvita, 2001 A Manual of Linguistic Field Work and Indian Language 
Structures, LINCOM Handbooks in Linguistics 17 (ISBN 3 89586 401 3)

There are some interesting notes on a comparativist's approach to what are 
and are not categorizable sound changes in:  
Wares, Alan 1968. A comparative study of Yuman consonantism. Janua Linguarum, 
series practica, 57. The Hague: Mouton. 

There's a lot more out there, particularly in Asia and Australian field work 
that I think MAY suggest to you that the steps or processes are arranged 
somewhat differently or rather more complexly than you have them, and that 
this may be more apparent in work with unwritten languages, and more opaque 
where writing is involved.

Steve Long