assignment 6#

This week, rather than answering questions about an article, you will be analyzing some child speech data for a few of the fundamental measures we use to understand language development! With this assignment, you will see two files, Adam_01 and Adam 20, two samples of speech taken from a typically developing child code-named “Adam”. Roger Brown and his students collected transcripts from Adam and two other children every week for three years in the Boston area in the early 60s. This is one of the first systematic longitudinal studies of children’s speech

Nowadays, these and hundreds of other transcripts are maintained in an online database (CHILDES) and most transcript analyses are done automatically. But the original analyses were done with pencil and paper, and you’ll be practicing that this week.

Brown found that the one of the best predictors of the grammatical complexity of children’s speech IS NOT their age but rather the mean length of their utterances.

For your pre-section assignment, please do the following (Detailed instructions are below, read them through before you start!):

• Read through all of Adam_01. Choose a place to start, and write down the first 15 utterances that Adam (not the other speakers!) makes from that point forward
• For each utterance, calculate the MLU (i.e. make a table like the one shown below).
• What is the average MLU in the section of Adam_01 you transcribed?
• Repeat this process for Adam_20. What is the MLU for this section?

• Pick one of the following grammatical morphemes:
  1. -ing (e.g., “they are laughing” vs “they are laugh” or “they laugh”)
  2. in and on (“put it in the box” vs “put it the box”)
  3. auxiliary “to be” (e.g., “he was swimming” vs “he swimming” or “he are swimming”). In these cases the form of be is not the main verb of the sentence. Often in these cases the auxiliary can be contracted (“He’s swimming” or “They’re dancing”).
  4. mandatory articles (“put it in the box” or “put it in a box” vs. “put it in box”)
• Starting at the beginning of Adam_20 find candidate cases where your morpheme should have been used (That is, compare what the child said to the grammatically proper way to say it – the “adult” comparison). Obviously to do this you have make a guess about what the child “intended” to say. Use context and the comment lines to help you do this.
• Make a table that shows what the child actually said, what they ‘should’ have said (even if it’s the exact same thing), and whether or not the morpheme was used correctly (i.e. mark all places where you morpheme should occur but is either omitted or replaced with an incorrect form.)
• Continue until you either find 20 contexts in which the morpheme should have appeared OR you reach the end of the transcript
• What percentage of the time does Adam use this morpheme correctly?

Part 1: Estimate the child’s mean length of utterance

To calculate MLU, follow these steps:

1. Identify the first 15 utterances from your starting point in each file that meet the following criteria. You’ll estimate MLU based on these 15 utterances (real MLU estimates use 100).  

• Utterance must be completely intelligible. If even one word in an utterance was not intelligible, that utterance is excluded from the analysis. Words that are unintelligible are transcribed as x.
• utterance is not a repetition (exact or partial copy of an immediately preceeding adult utterance)

2. Count the morphemes in each utterance.
3. Add the number of morphemes for all 15 utterances.
4. Divide by 15 to get the mean, or average.

How to Count Morphemes:

DO NOT COUNT

1. Do not count words which are false starts, reformulations, or repetitions unless the repetition is for emphasis. (e.g. “[then] then [he go] he went to the zoo” is counted as 6 morphemes; “No! No! No!” is counted as 3)
2. Compound words, reduplications, and proper names count as single words. (e.g. railroad, choo-choo, Big Bird)
3. Irregular past tense verbs and irregular plurals count as one morpheme. (e.g. took, went, geese, men)
4. Diminutives (e.g. doggie, horsie, dollie) and catenatives (e.g. gonna, wanna, hafta) count as one morpheme. It is assumed that the child understands these catenatives as single units, as opposed to understanding they are short for “going to,” “want to,” “have to,” etc.
5. Do not count fillers (e.g., um, well, oh).

DO COUNT

1. The –s plural marker (e.g. cat/s, apple/s). Count it even when used on irregular plurals (e.g. mouse/s). Exception: plurals never occurring in the singular (e.g. pants, shoes, clothes) count as just one morpheme.
2. The –ed past tense marker (walk/ed, count/ed). The –ed morpheme is counted even when used improperly (go/ed, drink/ed).
3. The –ing progressive tense marker (walk/ ing, count/ing).
4. The –s third person present tense marker (e.g. He like/s candy. Sue walk/s faster than Sara.) Exception: “does” counts as one morpheme.
5. Possessive –‘s marker (e.g. mom’s, boy’s)
6. Contractions (e.g. she’s, he’ll, they’re, what’s, she’d, we’ve, can’t, aren’t). Exceptions: “let’s,” “don’t”, and “won’t” are assumed to be understood as single units, rather than as a contraction of two words, so are just counted as one morpheme.

EXAMPLE: Calculate the MLU for the following utterances:

#	Child’s utterance	# of morphemes	Notes
1	I want ball.	3
2	That blue.	2
3	Give it to me.	4
4	Doggie’s hungry.	3	Doggie = 1 morpheme‘s (is) = 1 morpheme
5	Doggie eated .	3	eated = 2 morphemes
6	xx wagon.	–	utterance excluded because of unintelligible word
7	I hafta go potty.	4	hafta = 1 morpheme
8	She likes toys.	5	likes = 2 morphemestoys = 2 morphemes
9	(He go) he go bye-bye.	3	bye-bye = 1 morpheme“he go” is repeated, only counted once
10	The mice are sleeping.	5	mice = 1 morphemesleeping = 2 morphemes
	TOTAL	32

32 total morphemes divided by 9 utterances (since #6 was excluded) = a mean (average) of 3.6 morphemes per utterance, or an MLU of 3.6.

Part 2: The development of grammatical morphemes

Background: Roger Brown identified 14 morphemes that children acquiring standard American English tend to acquire in a similar order. These are not acquired in an all or none fashion. Children’s competency with using these morphemes changes over time. To chart development of these morphemes, we need to 1) find all the sentences where the morpheme should be used and 2) calculate the percent of times that the morpheme is accurately used in those contexts.

For example, in the sentence, “They are going to town,” the contractible auxiliary verb “are” was grammatically obligatory, or necessary. In other words, the sentence would be grammatically incorrect if a speaker of standard American English said, “They going to town.” Likewise, in “Mama bake cookies yesterday,” regular past tense marker “-ed” was obligatory, but in this case, left out. It is another example of a grammatical morpheme missing in an obligatory context.

We will be analyzing Adam’s session 20 (3;0). To calculate a child’s accuracy in using your morpheme, follow these steps:

5. Compare what the child said to the grammatically proper way to say it (the “adult” comparison). Obviously to do this you have make a guess about what the child “intended” to say. Use context and the comment lines to help you do this.
6. Please note that you really can’t analyze one word utterances in this way. If you cannot determine (with reasonable certainty) what the adult equivalent utterances is, do not include the utterance in your count
7. Mark all the correct uses of the target morpheme.
8. Mark all places where you morpheme should occur but is either omitted or replaced with an incorrect form.
9. Continue until you either find 20 contexts in which the morpheme should have appeared OR you reach the end of the transcript
10. Divide the number of correct uses of the target morpheme by the number of total times it should have been used (correct uses / correct uses + incorrect uses or omissions).