a bunch of bound morphemes floating in colored bubbles

The SALT reference databases use a very specific set of conventions for marking bound morphemes. Our protocols do not represent the only way to mark morphemes but if you want to compare your sample against our databases, it is important for you to use these same conventions.

If you aren’t interested in the nuts and bolts, just skip down to the “Summary” section. It’s the most condensed version of the SALT rules for marking bound morphemes that we can manage!

The SALT rules for marking bound morphemes may seem obscure at first. But there is method to our madness. This blog explains the rationale behind SALT’s conventions for marking bound morphemes. It goes beyond, “because it’s been this way for a long time and here are the rules to memorize”. Or, a little more clinically, “because these are the developmental morphemes and here are the rules to memorize”. Our hope in writing this blog is that by understanding the “why” behind the conventions, the rules will make more sense and, hopefully, it will be easier to implement them.

Pic of Roger Brown smiling in his office

Roger Brown
Harvard University

To a large extent, SALT’s convention for marking bound morphemes is patterned after the conventions used by Roger Brown (1973) in the calculation of MLUm (mean length of utterance in morphemes) back in the early days of language sample analysis.

Inflectional morphemes versus Derivational morphemes

Many of the explanations and examples in this section are taken from the following websites:
https://semanticsmorphology.weebly.com/inflectional-and-derivational-morphemes.html
http://www.mathcs.duq.edu/~packer/Courses/Psy598/Ling-Morphology.pdf

Brown excluded derivational morphemes in the calculation of MLUm. To understand why, we first need to distinguish between the two types of bound morphemes – inflectional and derivational.

Inflectional morphemes are used to show some aspects of the grammatical function of a word. They are always suffixes and always result in the same part of speech. We use inflectional morphemes to indicate if a word is singular or plural, whether it is a comparative or possessive form, and to mark tense. Inflectional morphemes never change the grammatical category(part of speech) of a word. For example, shoe and shoes are both nouns, tall and taller are both adjectives, and look and looked are both verbs. The inflectional morphemes simply produce different versions of the words.

There are eight inflectional morphemes. They are shown in the following table:

Inflectional Morphemes Added to Example
-s, -es plural nouns I have two black cats.
-‘s, -s’ possessive nouns My dog’s bark is very loud.
-er comparative adjectives I have longer hair than you do.
-est superlative adjectives He has the biggest pumpkin.
-s 3rd person singular verbs She runs fast.
-ed past tense verbs He played basketball.
-en past participle verbs She has eaten everything.
-ing progressive tense verbs He is playing basketball.

 

When children use inflectional morphemes, they are (generally) demonstrating their knowledge of the base word as well as their ability to encode the plural, possessive, or tense of that root word.

Derivational morphemes, in contrast, are used to create new words or to make words of a different grammatical class (part of speech) from the root form. For example, by adding the derivational morpheme -er the verb read becomes the noun reader. The addition of -ize changes the adjective normal to the verb normalize. Similarly, we can derive the adjectives helpful and helpless by adding -ful and ‑less to the noun help.

However, some derivational morphemes do not change the grammatical category of a word but they do significantly change the meaning of the word. For example, we can derive the nouns neighborhood and kingdom by adding the derivational suffixes ‑hood and ‑dom to the nouns neighbor and king. And derivational prefixes such as un- and re- generally do not change the category of the word to which they are attached. Thus, both happy and unhappy are adjectives, and both fill and refill are verbs. But each of these pairs of words, although clearly related, have very different meanings.

Derivational morphemes may be either suffixes or prefixes and usually, but not always, result in a different grammatical category. The following table lists some of the common derivational morphemes:

Common Derivational Morphemes (Suffixes) Added to Results in Examples
-ize nouns
adjectives
verbs
verbs
rubberize
normalize
-ful nouns adjectives playful, helpful, beautiful
-ly nouns
adjectives
adjectives
adverbs
manly, friendly
proudly
-sion verbs nouns discussion
-hood, -dom nouns <same> neighborhood, kingdom
Derivational suffixes which overlap with inflectional suffixes
(though they serve a different purpose)
-er verbs
nouns/verbs
nouns
nouns
reader
grader
-ed verbs adjectives am tired, was bored
-en verbs adjectives this spot is taken
-ing nouns/verbs
nouns/verbs
nouns
adjectives
biking is fun
interesting story
Common Derivational Morphemes (Prefixes) Added to Results in Examples
un-, a-
dis-, re-
anti-
Adjectives
verbs
nouns
<same> unhappy, atypical
dislike, refill, reevaluate, review
anti-aircraft

 

According to Brown, young children generally do not learn a base word and then apply a derivational morpheme to encode extra information. Instead, they usually learn these as fully-formed, independent words with their own specific meaning. Although derivational morphemes can logically be split into a root word and a prefix or suffix, these smaller parts are not meaningful to the child speaker and so they should not be considered separate morphemes in this case.

Guo, et al. (2018) gives evidence for this, including:

  • There is ample psycholinguistic evidence that base words and derived words (e.g., beauty, beautiful) are stored as separate lexical entries and should be given equal weight.
  • Children may learn a derived word (e.g., beautiful, interesting) before the base word (e.g., beauty, interest). It seems unlikely that children would add the derivational morphemes to these base words to form the derived words. Side note: to test this, we looked at samples from 355 typically-developing children under the age of 7 taken from the SALT Play and Conversation databases. The derived word beautiful was used five times while its base form beauty was only used once. And interesting was used twice and there were no instances of its base form interest.

They consider that derivation is a word-formation process, not a grammatical encoding process. Therefore, derivation reflects a speaker’s lexical skills, not grammatical skills.

So what are the rules behind the SALT conventions?

RULE 1: Do not mark derivational morphemes. Do mark (most) inflectional morphemes.

Why?

In short, we want to mark bound morphemes when they reflect the child speaker’s understanding that the prefix/suffix has a meaning separate from the root word. When using derivational morphemes – learned as fully-formed, independent words – a child speaker is only utilizing a single meaning. In contrast, when using an inflectional morpheme, the child is utilizing two meanings: the root word and the encoded plural/possessive/tense meaning.

By not marking derivational morphemes, we do not give the speaker credit for bound morphemes which change the meaning of the word (e.g., happy à unhappy) or change its grammatical category (e.g., friend à friendly).

However, when the child speaker is likely to have understood the separate meanings of the bound morphemes, we do want to mark them. So we mark most inflectional morphemes. Most, but not all…

RULE 1a: Do not mark the comparative (inflectional) morphemes -er and -est.

Why not?

Although -er and -est are inflectional morphemes, Brown did not count them because they are not obligatory. According to Guo, et al (2018), this means that it is a stylistic choice whether to use comparative and superlative form rather than the uninflected adjective. For instance, when given a choice of several balls of varying sizes, a child may select the largest one and say, “I have the big one” unless prompted to make a comparison.

RULE 1b: Do not mark irregular forms.

Why not?

Irregular forms are counted as single morphemes because children (generally) learn them as separate forms, rather than inflections of their base forms.

The following table lists examples of irregular words:

Category Examples of irregular words
plural man → men, foot → feet, cactus → cacti, deer → deer
all possessive pronouns I → mine, he → his, she → hers, we → ours,
you → yours, it → its, they → theirs
3rd person singular have → has, is → was
past tense begin → began, break → broke, go → went, get → got
past participle
(regular form is present tense + EN as separate syllable)
begin → begun, break → broken, go → gone,
get → gotten, see → seen, be → been
negation will → won’t

 

Some words are irregular because the sound of the base form changes. These words follow the standard spelling for inflected or contracted words but change its sound. Some examples follow:

Category Examples of changed sound
plural leaf → leaves, wolf → wolves
3rd person singular do → does, say → says
past participle drive → driven, write → written
negation do → don’t

 

RULE 1c: Do not mark plurals for words which do not have a singular form.

Why not?

Children would not have learned the singular form in order to then apply the rule for plurals. Following are some examples:

Examples of Plurals Without a Singular Form
belongings
binoculars
breeches
clothes
drawers
glasses (spectacles)
goggles
jitters
knickers
pajamas
pants
panties
remains
riches
shenanigans
shorts
suds
tights
trousers
tweezers

 

RULE 2: Do not mark concatenatives.

Why not?

Brown counted concatenatives as single morphemes because, like irregular forms, children may have stored them as holistic chunks. Following is a list of concatenatives:

Examples of Concatenatives (meaning)
betcha (bet you) liketa (like to) outta (out of) useta (used to)
coulda (could have) lookit (look at it) shoulda (should have) wanna (want to)
gonna (going to) musta (must have) sposta (supposed to) whatcha (what are you)
gotta (got to) oughta (ought to) trynta (trying to) woulda (would have)
hafta (have to)

 

RULE 3: Mark contracted words.

Why?

Contractions combine two words into one (e.g., we are à we’re). The speaker is given credit for the same number of morphemes whether using two words or the one contracted word.

Contractions Examples
-‘t
-n’t
negation I can’t leave yet.
He doesn’t know better
-‘s is
has
does
us
It’s time to go.
He’s been sick.
What’s he do for a living?
Let’s go.
-‘re are You’re late.
-‘m am I’m ready to take the test.
-‘ll will I’ll wait over here.
-‘d would
had
did
He’d do it.
He’d better leave now.
Why’d the boy look over there?
-‘ve have We’ve a lot to do.

 

Summary

These rules can be summarized as:

  • Only mark the following inflectional morphemes and contractions:
Inflectional Morphemes Contractions
/s plural /’t, /n’t negation
/z possessive /’s, /’re, /’m is, are, am
/3s 3rd person singular /’ll, /’d will, would
/ed past tense /’ve, /h’s, /h’d have, has, had
/en past participle /d’s, /d’d does, did
/ing progressive tense /’us us
  • Do not mark irregular forms, concatenatives, or plurals which do not have a singular form.

Formatting Notes

  • Use a slash (/) for bound morphemes which follow the free morpheme (suffixes) and use a backslash (\) for bound morphemes which precede the free morpheme (prefixes). There should be no spaces between the free morpheme and the bound morpheme(s).
  • When the spelling of a free morpheme such as CRY changes with the addition of the bound morpheme, use the root spelling of the free morpheme (as if the bound morpheme is not there). Then, simply add the slash plus the bound morpheme (i.e. CRY/ED). If this is not done, the stem CRI will be treated as a different word from CRY and thereby inflate Type-token ratio (TTR) as well as Number of Different Words (NDW).

If you would like to see some examples of the formatting for bound morphemes, see this PDF:
https://saltsoftware.com/media/wysiwyg/tranaids/TranConvSummary.pdf

Also, try our online course #1304: Transcription – Conventions Part 1:
https://saltsoftware.com/training/self-paced-online-training/course-1304-transcription-conventions-part-1

A Final Note

While we feel that Roger Brown’s research published in 1973 has stood the test of time, we understand that it is not the only way to understand the use of morphemes or to calculate MLUm. If you have a language sample which you want to compare with samples selected from the SALT reference databases, then these conventions are the most appropriate. However, these conventions may not be appropriate for all speakers or in all cases. Future blog posts will highlight some of these alternate scenarios and how they can be accommodated using custom coding schemes in SALT.

 

References

Brown, R. (1973). A first language: The early stages. Cambridge. MA: Harvard University Press.

Guo, L., Eisenberg S., Bernstein Ratner, N., & MacWhinney, B. (2018). Is Putting SUGAR (Sampling Utterances of Grammatical Analysis Revised) Into Language Sample Analysis a Good Thing? A Response to Pavelko and Owens (2017). Language, Speech, and Hearing Services in Schools, doi:10.1044/2018_LSHSS-17-0084.