SALT (5.0)

SALT stands for Systematic Analysis of Language Transcripts. It is designed to help you analyze and interpret language samples from one or more speakers during a communicative interaction. SALT provides clinicians and researchers with the means to transcribe language samples into a common format and to compute a series of general analyses of lexical, syntactic, semantic, pragmatic, rate, fluency, and error categories.

Collecting Language Samples

The SALT program will analyze language transcripts of any length. However, it is recommended that sample contain at least one hundred completed and intelligible verbal utterances of the target speaker.

Getting Started

1. Open the SALT program.

2. Click on the new file icon or select New from the File menu.

3. A dialog box will appear.

A. The first two lines allow you to identify the speakers in the transcript and to choose a target speaker. This is the only information that is required.

B. Fill in the child's Name, Female or Male, Date of Birth, and Date of Sample. The chronological age will be calculated for you.

C. Select Conversation, Narration, or Picture Description for the Sampling Context.

D. Enter the name of the examiner and the transcriber. This may be the same person.

E. Click on the Check button to check the header information for errors. Click OK to continue.

F. Click OK to exit the dialog box.

4. Move the cursor to the beginning of the blank line following the header information. You are now ready to begin transcription.

A. Begin each line of the transcript with the first initial of the examiner and subject's name.

B. All utterances must end with one of the following end of utterance punctuation marks:

1. Statements (.)period or (!)exclamation mark

2. Questions (?) question mark

3. Intonation Prompt (~) tilde

An intonation prompt is an open-ended utterance used to prompt another speaker for a response using rising intonation. It is not presented in the form of a question but more of a fill in the blank.

4. Abandoned Utterances (>) greater sign

If a speaker voluntarily stops in mid-utterance, end that utterance fragment with a greater sign and no period.

If a speaker abandons an utterance mid-word, end the part of the word you can discern followed by an asterisk (*) followed by the greater sign.

5. Interrupted Utterances (^) carrot

If a speaker is interrupted before completing an utterance, end the utterance fragment with a carrot and no period.

If a speaker is interrupted mid-word, end the part of the word you can discern followed by an asterisk (*) followed by the greater sign.

C. Each utterance should start on a new line with a speaker identifier as the first character and the ending punctuation mark as the last character. Do not put more than one utterance on a line.

D. Special character lines are the non-utterance entries. They begin with one of the following characters ($+-:;=) and do not require final punctuation.

E. If a special character line is continued on the following line, the new line must be indented two spaces.

F. Punctuation within utterances

1. Within Words * / + [ ] < >

The only punctuation marks that can be used within a word are asterisks for omissions, plus signs for Spanish clitics, slashes for bound morphemes, square brackets for word codes, and apostrophes.

2. Between Words , " ( ) { } [ ] < >

The only punctuation marks that can be used between words are commas, quotations, parentheses ( ) to mark mazes, braces { } for comments, square brackets [ ] for utterance codes, and angle brackets < > to mark overlap in talking between speakers. Do not use any other punctuation marks or symbols such as $, /, #, etc. unless they are serving as "codes" within square brackets [ ] or are contained in braces{ }.

3. Special Character Lines ($ = - : ; =)

You may use any symbols on special character lines as long as the conventions for special lines are not violated.

G. Bound Morpheme Conventions- If you want to calculate MLU in morphemes, Brown's Stage, Expected Age Range, and provide a Bound Morpheme Table, you must separate each bound morpheme from the free morpheme with a slash (/). Do not put spaces between the slash and the morpheme.

1. Possessive Inflections (-s, -'s)

Use Z so that DAD'S becomes DAD/Z. Note that words like HERS, OURS, and YOURS are considered entire units and are not slashed.

2. Plural Noun Inflections (-s, -es)

Use S so that BABIES becomes BABY/S and HOUSES becomes HOUSE/S. Do not mark nouns which only have a plural form, i.e. use PANTS and CLOTHES.

3. Plural & Possessive Inflections (-s')

Use S/Z so that BABIES' becomes BABY/S/Z and FATHERS' becomes FATHER/S/Z.

4. 3^rd Person Singular Verb Forms (-s, -es)

Use 3S (for both -s and -es forms) so that GOES becomes GO/3S. Note: Use DOES without a slash (considered one morpheme).

5. Other Verb Tense Inflections (-ed, -d, -ing)

Use ED (for both -ed, -d forms) so that LOVED becomes LOVE/ED. Use ING so that DOING becomes DO/ING and HAVE becomes HAVE/ING.

6. Contractible Verb Forms (-'m, -'s, -'ll, -'re, -'ve)

Use the free morpheme root spelling with the contracted verb stem so that I'M becomes I/'M.

7. Negative Contractions (-n't, -'t)

Use the root spelling of the free morpheme followed by /N'T or /'T. DOESN'T becomes DOES/N'T, CAN'T becomes CAN/'T, and DIDN'T becomes DID/N'T. Note: Use DON'T, WON'T, and AIN'T without a slash (considered one morpheme).

H. Mazes "( )" - Refers to a false start. 1. To mark a maze in an utterance you enclose it in parentheses.

2. Example- He gave me (the) the book.

3. Filled pause words such a "um" and "uh" may be marked as mazes.

4. If a speaker stutters or fails to complete a word, end the portion of the word you can discern followed by an asterisk (*). Part words usually occur in mazes (marked with parentheses). The only exception is when the speaker abandons an utterance or is interrupted mid-word.

I. Unintelligible Segments "X"

• Use X to designate each unintelligible word/syllable. You may type in as many X's as are needed.
• Words that just begin or end with X's are considered intelligible. (For example "XXer" "mXXX")
• A word can be made up entirely of X's.
• As a general rule, if you cannot understand a speaker's utterance after listening to it three times, consider it to b partly unintelligible or completely unintelligible.

J. Phonetically Consistent Forms "PCF"

PCF is an utterance or part of an utterance that fails to approximate the adult form and does not show consistent application to objects or situations but which tends to be stable in production with distinct prosody). For example: (PCF with embedded phrases)

$ Child, Examiner

E Can you see the baby?

C PCF { sounds like "tata" }

K. Idiosyncratic Forms "%word"

Type a % mark preceding a word which is used as an idiosyncratic form.

For example: C See my %vroom.

In this example, vroom means car.

L. Nonverbal Utterances: This is used to code a nonverbal turn.

Place a description of the nonverbal utterances as an embedded comment within braces { }. A nonverbal utterance is considered to be one utterance for calculation of speaker turn length.

For example: E Show me the ball.

C {points to the ball}

E Good job!

L. Overlapping Speech "<>"

Mark overlapping speech between speakers by enclosing th overlapping segments

in angle brackets <>.

For example: E Look at the <circus clown..

C <He has> a funny hat.

The examiner said "circus clown" at the same time as the child said "He has."

M. Spelling Conventions

1. Type in the exact words or glosses for each speaker.

2. Filled pauses such as "uh" and "um" are marked as mazes.

3. Numbers may be entered in written form or expressed as digits.

4. Hyphenated words should appear as one word.

5. When names or titles are used, you should give the speaker credit for just one word. For example: "Little_Red_Riding_Hood"

N. Other Transcript Entry Conventions (See Complete Instruction Manual)

1. Pauses p. 20

2. Timing Information p. 22

3. Omissions of Words p. 24

4. Root Identification p. 26

5. Codes p.27

6. Prompts and Encouragement p. 31

7. Transcriber Comments p. 32

8. Spanish Transcripts p. 34

5. Go to the Check Menu

The check menu is used to scan the current transcript for these transcript-entry errors and to help you correct them.

1. If errors are found in the transcript, a box appears listing the errors.

2. Select Correct Errors and the cursor will be positioned at the first error found in the transcript with the corresponding error message displayed at the bottom of the screen.

3. To locate the next transcription error that was found, select Next Error, to go back to the previous error, select Previous Error, select First Error to go to the first error, and select Last Error.

4. When you reach the end of the errors, you will be asked "Would you like to re-check your transcript?" Select Yes.

5. The error checking routine looks for errors in format and punctuation. It does not recognize misspelled words. Once you have corrected all errors, you may run the transcript through "Word Root Table" and "Bound Morpheme Table" in the Analyze Menu looking for any "unusual" words and bound morphemes.

6. The Analyze menu contains almost all the available analyses. When you select Analyze from the main menu, you are presented with a menu of the available analyses.

6. The Settings Menu contains the following:

a. Analysis Set: This is used to specify the target set of utterances that you want to base the analyses on.

b. Transcript Cut: This determines how much of the transcript to include in the analyses.

c. Data Directory: This is used to change the current data directory to the folder containing your data.

d. Standard Word Lists: This is used to edit the contents of the standard word lists.

e. Language: This is used to change the current language setting.

7. The following are analyses offered:

a. The Transcript Summary gives you an overall view of the transcript cut.

b. The Word and Morpheme Summary gives you an overall summary of the words and morphemes found in the transcript cut. It includes: total words, different words, TTR, MLU, and Brown's Stage. All values are calculated for both speakers and are given separately for the current analysis set and for the total utterances.

c. The Discourse Summary provides you with information on the interaction between the speakers by analyzing the types of utterances used.

d. The Turn Length Summary contains information on speaker turns lengths, measured in utterances and words.

e. Rate and Pause Summary

f. The Maze Summary gives an in-depth analysis of the number and types of maxes in the transcript cut.

g. The Code Summary lists and counts all the codes in the transcript cut.

h. Utterance and Maze Distribution Tables

i. Word Root Tables

j. Bound Morpheme Tables

k. Word Code Tables

l. Utterance Code Tables

m. Standard Word Lists

n. Standard Utterance Lists

This is only a summary of the manual. If you have any questions refer to the manual for more in-depth information. Salt is installed on computers in rooms 205 and 217 in the clinic.