transcription_guidelines.txt

BeRP Transcription Guidelines:

Pauses which are longer than 1 sec get the dot.
Use only one, even if the pause is ten seconds long.

.

Filled pauses: We only use these four
 [er]
 [mm]
 [uh]
 [um]

Other noises:
 [beep]
 [laughter]
 [lip_smack]
 [loud_breath]
 [noise]
 [tap]
 [unintelligible]   (this is in case they say something you can't decipher)

Everything is lower case.
 i
 chuck

Restaurant names:
 Use spaces in them when the name consists of multiple
	english words, or names in which only one word is foreign:

	meal ticket
	won thai cuisine
    gertie's chesapeake bay cafe
	bette's ocean view diner

Names which are multiple foreign words:
should not have spaces in them; use dashes instead of spaces:

 	a-la-carte
 	au-coquelet

Exceptions to this rule are names where people refer to
the restaurant by individual parts of the name (such as below,
where people called it 'eiffel' or 'tour' as well):

 	la tour eiffel

Street/city names which have two or more English words: use spaces;
    shattuck avenue
	martin luther king

Street/city names which have two or more foreign words: Use __:
 	san__pablo
 	san__francisco


Words which consist of letters of the alphabet:
Add two underlines "--" between the letters.
(This is different than MADCOW standard, they would
say "a m" instead)

 a__m
 p__m

Genitives, other apostrophe-s, other apostrophes:  DO include the apostrophe:
 augusta's
 chuck's
 that's
 christopher's-cafe
 i'd
 i'll
 i'm
 i've
 how's
 o'clock

The name of this place:
 i__c__s__i (If spelled out)
 icksee (if pronounced as a word)

Contractions:  Use your judgement, but DO use the obvious ones:
 wanna
 gotta
 hafta

THe word "OK", don't use "ok", use:
 okay

This is a word: 
 oh


Fragments:
  words which are cut off, either because the speaker stopped
	in the middle of a word, or because the microphone turned
	on or off in the middle of a word, are marked with a dash "-":
	
	i wa- wanna  eat

	-o you wanna eat?

	Put the dash at the very beginning or end of a word, 

Mispronunciations:  if they mispronounce a word (not just
	a dialectal variation, but a true mistake).  For example,
	if they prononce 'understand' as "unshtam" or something weird
	like that, spell the word with its standard orthography,
	but put asterisks on either side:

	i don't *understand* 

Angle brackets:  used for verbal repairs, to mean that
   the "correct" parse of the sentence should ignore these words.

  <i> i want to eat now.

   if there is more than one word in the repair, put
	angle brackets around EACH word:

	<i> <want> <to> [uh] i wanna eat some chinese food

Parenthesis:  Used for word-fragments, to indicate that the
	part of the word in parenthesis is the transcribers
	BEST GUESS as to what the speaker meant to say.
	(this usually comes up when the speaker started
	before the mike was turned on).  Thus in the following,	
	we THINK the speaker meant to say 'this', but the only
	thing we actually have phonetically in the file is the 's'.

           (thi)-s is for breakfast

Squiggly brackets:  I forget what these mean, I'll remember at some point.