Techniques
Techniques
Techniques
Basic techniques:
• Start with 2 forms that minimally differ
• Whatever differs both phonologically & semantically is likely
a morpheme
• Caveat: a morpheme with 2 different forms depending on
context (i.e. allomorphy, such as English plurals –s, -es, -en)
• Factor one hypothesized morpheme out of the data set
• Move on to a 2nd pair hypothesizing a 2nd morpheme
• Whenever possible, find a word containing both morphemes
• Continue this process until each morpheme in every word is
accounted for
• Note any generalizations (recurring patterns) wherever
supported by the data
1
Practice dataset 2: Ga~
mi'fo'kø` 'I have not wept' mi'fo˜ 'I will not weep'
ebaa'du' 'he will cultivate' wøbaa'la' 'we will sing'
obaa'la' 'you will sing' eyo'o'kø` 'he has not recognized'
mi'du'kø` 'I have not cultivated' obaa'hao 'you will worry'
ma'du' 'I will cultivate' ela'˜ 'he will not sing'
wøfo'˜ 'we will not weep' odu'˜ 'you will not cultivate'
ola'kø` 'you have not sung' ma'fo' 'I will weep'
2
Practice dataset 4: -en in English