Root Identification

Marking Roots

Be sure that the vertical bar and the root word directly follow the word used with no spaces between. For example,

In the first example, the child says "He wented home." The utterance is transcribed as it was spoken. The extraneous bound morpheme "ed" in the word "wented" is NOT slashed. "Wented" is then defined as a form of the word "went". This is done to avoid inflating the calculation of "number of different words". "Wented" is counted as another, albeit incorrect, instance of the word "went". The other two examples are marked as such for the same reasons as the first example.

Marking Overgeneralizations as Errors

Overgeneralized words, as in those in the previous examples, may also be coded as errors. The standard error code [EW] is attached to the end of each word. For example,

Root Identification and Irregular Verbs

When transcribing English language samples for the SALT reference databases, the only words that are identified are the overgeneralization errors described previously. If you choose, however, you can mark all irregular past tense verbs with the present tense word root so that all tenses of the verb are recognized as belonging to the same word root. For example,
   C Go away.
   C The frog went|go away.
   C The next day he was gone|go.
In this example the verbs "went" and "gone" are both identified as forms of the word root "go". This identification affects the calculation of "number of different words" which, in this example, decreases from 10 to 8.

If you wish to give the speaker credit for the past tense morpheme when using irregular verbs, add the inflection to the redefined root. Including the inflection when redefining irregular words affects calculations such as "mean length of utterance in morphemes". For example,
   C The frog went|go/ed away.
   C The boy got|get/ed his feet wet.
In this example the verbs "went" and "gone" are both identified as forms of the word root "go". Including the inflection when redefining irregular words affects calculations such as "mean length of utterance in morphemes" which, in this example, increases from 4.5 to 5.5. The calculation of "mean length of utterance in words", however, is not affected.

Identifying irregular words and adding inflections affect the results of the analyses. If you choose to code your transcripts using these conventions, be careful if you compare your transcripts to transcripts selected from the SALT reference databases. When transcribing English language samples for the SALT reference databases, the only words that are identified are the overgeneralization errors.