Stem morpheme identification” as a means to suffix segmentation

We have concluded so far that instances of final -al, -ic, -ly, and -y are incapable of reliable identification with morphemes (AL), (IC), (LY) and (Y) owing to their relative lack of phonemic, graphemic and semantic distinctiveness.

We must approach the problem from a new direction. If we cannot identify -al, -ic, -ly and -y with the morphemes (AL), (IC), (LY) and (Y) by holding each one up for examination, we can at least aim atsegmenting the suffix by establishing the morphemic status of the rest of the word. Segmentation can be done in two ways:

S 1(= Segmentation rule no.1)

If we can recognize the stem (or that part adjacent to the affix) as appearing elsewhere in Webster or Brown as a free form we thereby establish the morphemic status of that form. By this means we isolate and segment the affix.

Example 1: Example 2:

fishy to be segmented; grandfatherly to be segmented;

fish appears in Webster as a free form; grandfather appears in Webster as a free form;

segmentation of fishy = fish- + -y; segmentation of grandfatherly = grandfather- + -ly;

Example 3:

unmanly to be segmented;

man appears in Webster as a free form;

segmentation of unmanly = un- + -man- + -ly.

S 2

If the stem or its final part doesnot appear elsewhere as a free form, the morphemic status of the stem can be proved by means of this pattern:

undeserv-: undeceiv- + -ed: -ing

(The shift within each of the two pairs must be semantically, phonologically, and graphemically consistent.) We then go on to segment the suffix. (These Segmentation criteria are modifications of rules presented by Ellegard, pp 7–11. Compare also Ljung (1974) page 15, rule “G 4”.)

Note that in both S 1 and S 2 we first establish the morphemic status of the stem, thereby isolating and segmenting the affix.

These two methods are adequate for most cases, but a smallish number of cases remain when we feel strongly that we have a right to segment but are not permitted to do so by rules S 1 and S 2. Such a case is cordial, where the stem cordi- is neither a free form nor segmentable by the pattern:

cordi-: ? + -al: ?

We have however the related form cordiform where we recognize form as ap­pearing elsewhere as a free form, leaving us witha boundform cordi- with morphemic status. With the help of this cordi- we can isolate and segment the suffix -al.

We need a new rule to enable us to use this principle for segmenting cordial.

This rule is S 3.

S 3

A stem can be segmented from its affix(es) if the stemor its affix-adjacent morpheme(s) appear in Websteror elsewhere in Brown asa boundform segmentable by S 1and/orS 2.

Date: 2016-04-22; view: 766