ACLEW FAQ
What do these terminology mean?
ACLEW: analyzing child language experiences around the world = this overall project
DARCLE: daylong audio recording of children's linguistic environments
DAS: darcle annotation scheme = this annotation scheme
AAS: aclew annotation scheme = builds upon the DAS, standardized for the needs of the ACLEW group
dependent tiers for main child
vcm - vocal maturity (L, Y, N, C, U)
lex - whether it contains lexical words (W, 0)
mwu - multiword utterance (M, 1)
dependent tier for everyone else
xds - addressee (A, C, B, P, O, U)
minCHAT: used for transcribing lines of speech; transcription must take place for all speaker utterances
Closed vocabulary - the multiple choice options for vcm, lex, mwu, xds
Open text entry - for orthographic transcription
Which template should I use with my file?
See tutorial 2 slide 6: based on the target child's age, select the respective template. For the high volubility project, the .eaf already has the template loaded in, and tells us which segments to code.
Lots of people talk in this file. Do I need to segment the speech for all of these people?
Segmenting electronic speech
Tutorial 4, slides 3 and 4
Q: Do we note 'MULT' if there are multiple sources of EE1?
I can't tell the difference between two speakers
Do I need to transcribe {adult, child, target child} speech?
Transcription should be done using minCHAT standards
What is the workflow?
Remember to save frequently, and have auto-save set as well!
See slide 11 of tutorial 3
Who are the addressees?
See tutorials
Note that all pre-recorded speech segmented on the EE1 tier should be coded as “O”. Even though this speech has been identified as directed toward children, its pre-recorded nature makes it slightly different than live speech.
What counts as a word?
hedges and expressives {um, oh, mm, ... } = each of these is one word as long as we can infer from context that they have a meaning
these mainly occur in speech from adults or from older children
in infant speech these sounds are coded as canonical or non-canonical babbling
determiners {a, the} = each of these is one word
proper names = one word
part words = one word, as long as it's clear which is the intended word
names of letters = word
consecutive names of letters should be a multi-word utt
sung words = words
What if two codes apply?
For example, if the child is laughing and canonically babbling at the same time
We should assign the value that approximates speech the most
If laughing (L) and canonically babbling (C), then C
If canonically babbling (C) and non-canonically babbling (N) in same segment, then C
see Tutorial 5, slide 7
If an utterance has a mix of canonical and non-canonical syllables, it should be labeled as canonical
NOTE: Annotators should always take the most speech-advanced label in cases where multiple labels could apply. If laughing or crying is present alongside C/N babbling, this should be noted in the “Notes” tier.
C > N > L/Y
What counts as a separate utterance?
For example, if someone says "did you take your socks off last night I think you did" with almost no pause, would the two sentences be separate utterances or one?
two separate utterances
Troubleshooting -- Dropdown menu for vcm/xds tiers is not working
If you don't get the choices for the vcm or xds tiers, go to Edit -> Edit Controlled Vocabularies. In the window that opens, select External CV and then try pasting the entire link below into the URL box:
If that doesn't work, hit Browse and navigate to Fas-Phyc-PEB-Lab/Seedlings/ACLEW and then select the "ACLEW-basic-vocabularies.ecv" file.
ELAN will tell you that "A Controlled Vocabulary with that name already exists" for each tier. Select Replace Existing CV each time.
Troubleshooting tip: ELAN keeps crashing
New Mac processors (e.g. produced after November 2020) do not interact well with ELAN-- look for ELAN update or use older machine, if possible.
Last updated