Documenting the Skolt Sami Rule file

Introduction

xfst vs. twolc

The Skolt Sami rule file differs from the other languages in being an xfst file, rather than a twolc file (each format have different chapters in the pdf version of the Xerox book). The differences between thesew two formats may be summarised as follows:

xfst rules are ordered, whereas twolc rules apply simultaneously
xfst rules stand in a feeding-bleeding relation to each other, whereas twolc rules do not
xfst rules are good at doing many-to-many replacement operations (also of different-size strings), whereas in twolc one must have one rule for each pair of members in equal-length strings
xfst cannot generalise over ordered sets of pairs, thus the Cx:Cy replacement of members of ordered sets Cx and Cy that is found in twolc must be done in a more clumsy way in xfst.

The reason why xfst is chosen for Skolt Sami (and in the future also for Inari Sami(?)) is that there are several cases of many-to-many replacements in the Skolt Sami grammar.

The structure of the rule file itself

The rule file for Skolt Sami is divided into 5 main types:

Definitions
Consonant shift rules (tbw)
Vowel alternation rules
Consonant gradaton rules
Rules for cleaning up and composing end result

The structure of this document

The next section presents a general discussion on the rule format of xfst. Thereafter, the rule types of the rule file will be dealt with, in the order indicated in the 5-point list given above.

Rule format and the cascade of rules

Xfst rules are of the following format: interface offers 4 different operator

The definitions

There is no section corresponding to the Alphabet section of twolc. The sets are defined as such by means of the operator OR (written "|" ).

The consonant shift rules

This section will include a.o. word-final consonant changes

The vowel alternation rules

Last modified: Tue Mar 4 12:35:30 GMT 2003