The Semantic Elaboration Coding System
All materials related to the SECS are copyright 2004 by John C. Thorne. All rights reserved.

Training Manual

A system for analyzing elaboration in narratives

+Describing a Scene

Stories are made up of events that happen to participants. To convey information about these events, language provides us with two basic structures. These structures allow us to talk about what happened, and what or who participated in this happening. In English, the talk about the happening is done via what we will call a “verb complex,” while the talk about the participants is done via what we will call a “nominal complex.” Working together, these complexes allow us to specify to varying degrees, from varying perspectives, with varying emphasis what happened and who or what was involved (Langacker, 1991). Manipulation of the elements in these verb complexes and nominal complexes changes the amount of information that is available to listeners for interpretation of the message, and their subsequent understanding of the event. Good story tellers will utilize these tools more effectively than poor story tellers and provide listeners who are unfamiliar with the content of the story with more information about the event with which they can, presumably, develop a better understanding of the event being described. The Semantic Elaboration Coding System (SECS) is an attempt to help quantify this manipulation and its impact on the amount of information conveyed by storytellers.

An example will help to illustrate this. Imagine a picture of a boy and a dog looking at a large frog sitting in a jar. The jar is at the foot of the boy’s four-post bed in his bedroom. The boy sits on a short footstool, elbows on his knees, in front of a large pair of black leather boots. Next to the boots on the boy’s left is a striped short-sleeve shirt. Over the bed hangs a cone shaped light. On the wall to the left of the bed is a slightly opened window through which a crescent moon is visible in the night sky. The dog is standing with his front paws and nose on the lid of the frog’s jar, with his tail wagging behind.

A person looking at this picture has many choices when deciding what to say to a person that cannot see this picture if they wish to convey an image or concept of what the picture is about. They can be very detailed and elaborate in their descriptions or they can be more schematic (with the above paragraph being an example of a somewhat, although certainly not maximally, elaborated example). Let’s look as some possible utterances that could be used to describe the picture.

S1). “A boy and a dog”

S2).”There is a boy and a dog and a frog”

S3) “The little boy has captured a frog and is showing it to his puppy before he goes to bed.”

S4) “Before going to sleep, Timmy, a young boy with dark brown hair, is sitting on a stool spending some time alone in his bedroom with his two pets, a little wiener-dog named Duke and a big green frog he keeps in a jar.”

Notice that both the participants in these events, and the events themselves are presented with differing perspective and with differing degrees of elaboration in the various sentences. The story’s protagonist is described as a boy (S1), a little boy (S2), someone who captures frogs (S3), a young boy with dark brown hair (S4), and a pet owner (S3, S4). The participants are described as simply existing together in a room (S2), as being displayed after capture (S3), and as spending time in communion before going to sleep (S3, S4).

When story tellers make particular choices in characterizing scenes depicted in the pictures, they provide their listener with differing levels of detail from the very schematic (e.g., S1) to the very elaborated (e.g., S4). It is unlikely that any two story tellers will make the exact same choices regarding how to describe the scene, but it is likely that the more skilled story tellers will provide more elaborated images of the scene than less skilled story tellers for most scenes in the story. They will do this by using more elaborated vocabulary clustered together into elaborated verb and nominal complexes that allow for richer interpretation of the scene by the listener.

The SECS described here is an attempt to quantify this process of elaboration by codifying the contents of the verb and nominal complexes storytellers use to convey information about a scene to their listeners.

Verb Complexes

For this analysis, the verb complex is defined as a verb root together with the satellites that help to refine its meaning. A verb root is a word that indicates action, state of being, or change of state which might stand alone (left, escaped) or work in conjunction with affixes or other free-standing words (went away, climbed out) to convey information about what happens in an event or scene or sub-scene. The verb root is what Croft (2001) calls the “primary information bearing unit” (PIBU) of the verb complex. This PIBU will exist alongside satellites to specify what happens in a scene. A satellite is any word or affix that works in conjunction with the verb root to expand, refine, construe, or constrain the meaning of the verb complex. For this analysis, affixed satellites (such as the “ed” in escaped) will be counted as part of the verb and will not be broken away from the root for separate analysis. Free-standing satellites will be identified for coding to recognize their impact on the meaning of the verb complex. The concept of a verb satellite comes from Talmy (2000: pg. 102) who points out that the boundaries of the class “verb satellite” are not yet clearly defined in theory. In his analysis, Talmy cites verb particles in English as a prototype for verb satellites but emphasizes that English has other hybrid elements that serve a similar function. For this analysis, any word or group of words that can be said to expand, refine, construe, or constrain the meaning of the verb complex and its relationship to other elements in the utterance will be included. So, in addition to verb particles, in this analysis satellites will include words and word groups traditionally described using terms such as auxiliary verbs, adverbs, predicate adjectives, conjunctions, prepositions, adverbial phrases and clauses, and other multiword constructions not included in Talmy’s characterization of the class “verb satellite.”

Verb roots and their free-standing satellites (working together as a verb complex) can indicate the fact of motion or the manner of motion, the fact of being or the manner of being, the fact of state change or the manner of state change. As well, the verb complex can indicate the direction, path, or placement of motion or change (stood or fell or went up or blew away or looked around), and/or the degree or intensity of motion, being, or change of state (i.e. whispered or yelled very loud or kinda pushing or slept lightly or crashed loudly), and/or other aspects of the event being described. Three codes will be used in this system to distinguish between these various functions of words in the verb complex: [vfact], [vmanner], and [vsat]. Each will be discussed in detail below.

+Verb Roots

Verb roots are the primary information bearing unit (PIBU) of the verb complex and are coded according to the level of elaboration they convey. Two levels of elaboration will be recognized: schematic verbs [vfact] and elaborating verbs [vmanner].

The first level of elaboration recognized are verb roots that tell the fact that something happened WITHOUT elaborating on how it occurred. These words will be coded as [vfact]. They describe an event in the most schematic of terms. If elaboration beyond the schematic level is present, and a schematic verb is used, the verb complex will also contain verb satellites that carry this information. The most commonly occurring schematic verbs are listed in table 1.

Table 1: Common schematic verbs [vfact].

Happen Exist Change
Go, went, going, goes
come, coming, came
move, moved, moving, leave, leaving, left, stayed
Is, was, were, are Died, die,
Stop, stops, stopping, stopped
Start, starting, starts
Got, get, gets,
have, had, has,
take, taking, took,
gave, give, gives
bring, brought, bringing
will be, being, be broke, breaking
Told, tell, telling,
said, say, saying,
Getting, got, gotten,
Do, did, doing, does Change, changing, changed
Make, made, making,
Try, trying, tried,
Put, putting, stay, staying
Work, worked, use, used
let, letting,
Happen, happening, happens, happened

To identify these schematic verbs determine if the verb root only indicates:

1. the fact that something happens/ happened/ will happen.
He went outside.
The dog took it.
He told him something.
Something happened to the frog.
2. the fact that something exists/existed/ is/will be in a state.
There is a boy and a dog.
The frog was not there.
He is being bad.
3. the fact that something changes/ changed/ will change state.
The jar broke.
The frog died.
He got/will get/ is getting mad/out/wet/red in the face.

If so they are coded as [vfact].

The second level of elaboration recognized are verb roots that not only indicate the fact that something happened, but also indicate how something happened. These verb roots are more elaborate in their characterization of the event in that they provide information regarding the manner in which something happens, exists, or changes. These elaborating verbs are coded as [vmanner]. Like schematic verbs, these elaborating verbs can stand alone (ran) or may work with verb satellites (ran through) to further elaborate the description of the event.

To identify them determine if the verb indicates:

1. the manner in which something happens/ happened/ will happen.

He walked/ran/strolled down the road” (compare to “went”)
His frog escaped/tip toed from the jar. (compare to “got out”)
He jumped on the rock. (compare to “got on”)
He tripped/fell/stumbled. (compare to “He went to the ground.”)
He grabbed/climbed a branch. (compare to “got/ got on”)
He searched/looked for his frog. (compare to “He went for his frog.”)
He remembered it.
He lost his frog. (compare to “did not have”)
He yelled/whispered/screamed. (compare to “told”)

2. Indicate the manner in which something exists/ existed/ is/will be in a state
He lived with a boy and his dog. (compare to “is“)
They slept while the frog escaped. (compare to “were there”)
The boy was worrying about his frog. (compare to “didn’t have”)
He seems angry. (compare to “He is angry”)

3. Indicate the manner in which something changes/ changed/ will change state.

Then the jar shattered. (compare to “broke”)
The dog was nearly killed by the fall. (compare to “The dog almost died”)
He woke up. (compare to “got up”)
The hive was crushed. (compare to “was in pieces”)

If so they are coded as [vmanner].

Verb roots will sometimes be composed of multiword idioms (as in tip toes or loopty looped).
In these rare cases, the complete idiomatic verb root will be coded as if it were a single word (tip toes[vmanner], loopty looped[vmanner]).

+ Verb Satellites

Verb satellites will be identified for their contribution towards elaborating the verb root’s meaning. They will be coded as [vsat].

To identify them, find words or word groups that elaborate the verb root. Some examples of the types of information that satellites can contribute to the verb complex are presented on the following page in table 2. Highlighted verb roots are presented in bold while satellites being highlighted are in Italics.

If words help to elaborate the verb complex they are satellites and are coded as [vsat]

Note that a verb complex may have many satellites (as in “Suddenly, the deer angrily threw them down off of the cliff.”) and that they are not necessarily clustered directly around the verb, but may be separated from the verb by other elements in the utterance. As well, recognize that an utterance may contain more than one verb complex (as in The boy ran and the dog barked at the bees).

Note that sometimes verb-roots are implied rather than explicit, but implied verb-roots can still have verb satellites. For example a child might say “They thought they saw the frog. No frog, just a gopher.” In this case, “no” and “just” are modifying an implied “There was,” In cases such as these, the satellites are coded [vsat] as if the verb were stated. If it is unclear what verb might be implied in the statement, leave the satellites uncoded (see below for guidelines on poorly formed statements.

+ Auxiliary Verbs

The verbs is/was/were/are/will be are often used to convey tense information as in “He was looking.” They can also be used to convey the fact that something exists as in “The boy was there.” In both cases, they are coded as [vfact]. The important thing to pay attention to is how the word that follows these verbs is coded. If the word is a verb such as “being, running, looking, searching” it will be coded as a verb. Like these examples:

He is[vfact] being[vfact] silly[vsat].
The dog was[vfact] running[vmanner].
They were[vfact] looking[vmanner].
They are[vfact] falling[vmanner] down[vsat].
He will[vfact] be[vfact] searching [vmanner] for the frog all day.

When these verbs convey existence they are followed by a satellite.

When they convey tense information they are followed by a verb.

This distinction falls on a continuum from very satellite-like to very verb-like as seen below:

He is[vfact] brown[vsat].
He is[vfact] mad[vsat].
He is[vfact] gone[vsat].
He is[vfact] missing[vmanner].
He is[vfact] running[vmanner].

To help make this distinction, look for “ing” verbs. These will be coded as verbs, either [vfact] (being) or [vmanner] (looking, running, searching).  Marginal cases such as “He was about to fall” can be tricky. The distinction is between whether the is/was/were/are/will be verb is telling you when something happened and when it is telling that something exists in a particular state.  So “He was about to fall,” which tells you the state “He” was in would be coded as follows:

He was[vfact] about[vsat] to[vsat] fall[vmanner].

Compare this to “He was falling” or “He fell.”

He was[vfact] falling[vmanner].
He fell[vmanner].

These verbs are frequently part of contractions such as “wasn’t, isn’t, weren’t, aren’t. Be sure to code both pieces of the contraction.
Like this: wasn’t[vfact][vsat].

Table 2. Common types of verb satellites [vsat]

Highlighted verb roots are presented in bold while satellites being highlighted are in Italics.

Direction or Placement or Path Manner or Mode
He found out/ stood up/ fell down.The boy took it away.He looked everywhere/ didn”t know where else to look. He went to bed.Keep out of there.Here it is/ It was here last night/ Where is it?

No matter where/wherever he looked it was gone.

The bees chased after/buzzed around the dog.

It is in the hole.

There was nothing there.

It sounded like there were frogs nearby.

He wanted to know if the frog was in there.

It was broken and crashed all over the ground.

He ran in/ It ran out.

He climbed on/ They stepped off (the cliff).

The bees drove off (the dog).

The dog leapt aside.

The owl came forth/The boy walked away.

They went ahead.

The frog never came back.

The dog ran across/ along.

They ran through/ It ran past/by.

The dog came over.

The boy spun around.

They walked around.

They walked (all) about.

The owl flew up.

The owl flew down.

They went above.

The dog went below.

A deer ran up.

The dog followed along after (the boy).

They slammed together.

They rolled apart.

It shrank in/ spread out.

There was a hole right in front of him.

The log was sitting right beside of him.

He ran fast/ran away quickly.He kept looking.He didn’t have his frog.He couldn’t find his frog.He may want to keep out of there.

He would like to know where his frog went.

The bees chased him angrily.

He slowly began to realize his frog was gone.

He thought he would look for him in the woods.

He seems/is/got mad/out/wet/red in the face.

He seems angry/ was gone/ does wonder.

The jar was too heavy on his head.

He was more worried than angry.

The jar wouldn’t pull loose.

The frog broke free (from the jar).

The dog stayed clear (of the swarming bees).

His tongue froze stuck (with fear).

The boy held fast (to the antlers).

They started over from the beginning.

They slept on all night.

They drove on from there.

The deer barged on in.

They worked away on the room.

He was running along when he tripped.

He climbed onto/into/inside of it.He looked for/worried about his frog.While he was running he saw it.After he got up he realized it was gone.Suddenly a deer appeared.

Then he went to sleep.

He was more worried than angry.

So, he gave one to the boy.

He slipped and fell.

The dog fell but didn’t hurt himself.

It sounded like frogs nearby.

He saw that the frog was gone.

The next morning he saw that it was gone.

He saw that it was gone and the dog didn’t.

He looked to see if the frog was in the hole.

He ran because the bees were mad.

The jar was too heavy on his head.

He thought he would look for him in the woods.

He used to be in the jar.

They burned up 2 hours.

The dog chewed up the boot in 10 minutes.

The dog teased the bees, so the bees got him back.

The deer ran with the boy on top of his head.

There beside him was a log making funny noise.

Degree or Intensity
He looked harder/longer/some more, but still couldn’t find him. No frog, just a gopher.He was more worried than angry.No matter how hard he looked…He looked really mad.He did not come at all.

It was so much bigger than he was.

The jar was too heavy on his head.

There was a log right in front of him

The log was sitting sorta beside him.

He was totally bummed out.

Highlighted verb roots are presented in bold while satellites being highlighted are in Italics.

Nominal Complexes

Analogous to the verb complex, which conveys information regarding the action/state-of-being/ change of state aspects of events, is the nominal complex, which conveys information regarding identity and properties of objects and participants in those events. The nominal complex is made up of a nominal root (a noun, nominal verb form, compounded noun) and its associated modifiers. A unique feature of the nominal complex involves the fact that the same objects/participants can participate in several events. This means that the person speaking can refine and add to the detail regarding a referent across several nominal complexes with each new reference either containing new information, or simply pointing to the previously indicated object/participant without adding any features. Recognizing this characteristic of nominal complexes, the SECS will track the objects/participants across the entire narrative in order to count any additional features contained in the various references to them.

+ Nominal Roots

Nominal roots are the PIBU of the nominal complex and are coded according to the level of information they contain much like the verb roots discussed above. The system will recognize relatively schematic nominal roots [ngeneral] and elaborating nominal roots [nspecific].

1. On the first reference to an object or character (or nominalized event/process) the nominal root is coded according to the specificity of its reference. So words with little specificity such as “person,” or “animal,” or “room,” or “tree,” or “frog,” or “dog,” or “All was lost” or “everything” or “stuff” or “thing” are coded as [ngeneral]. Nouns that contain more features of the object or character, such as “Timmy,” or “puppy,” or “terrier” or “oak,” or “muledeer,” or “hive,” or “bedroom” or “his barking didn’t help” are coded as [nspecific]. Notice that proper names and nominals referring to events/processes are coded as [nspecific]. See Table 3 for more examples likely to occur in the frog story.

2. A second (or subsequent) reference to a specific object using the same word is no longer coded according to the general/ specific distinction as the word simply refers to a previously identified object/character without conveying additional information.  This is coded as [nref]. This is true if the word is a second reference to a whole character or object (like “The boy” or “the jar”) or if it is a second reference to a part of that character or object (like “his head/horns” or “the point”), or if it is a second reference to a group of objects (“the frogs” or “the bees”). Shortened forms (such as “hive” for “beehive”), when used to refer to something that has already been referred to using the long form, are coded as [nref].

3. If a new word is used to reference something in the story, it is coded according to its specificity. If it is specific it will add additional information not contained in the initial form used for that object/character. For example, in “The kid, Timmy, is looking for his frog” the code for “Timmy” is [nspecific] since it adds naming information. All subsequent uses of Timmy or kid would be coded as [nref]. Other examples include “Something is in there. It’s a gopher” or “He grabbed the branches. But they were really a deer’s antlers.”

It is possible that the new word form is more general and does not add additional information. In these cases the word will be coded as [ngeneral]. An example would be “A gopher or something bit him on the nose.”  In this case “gopher” is specific and gets the code [nspecific] while “something,” which refers to the same thing is general, and does not add information about the thing, so it is coded as [ngeneral]. In all subsequent occurrences referring to that same animal, “gopher” or “something” (as in “that something that bit him”) would be coded as [nref].

4. Also notice that groups of words can be used in a repeated manner to refer to the same object or character. For example “the little boy” may be used each time the character is referred to. In this case the whole group of words is working as a unit to reference a previously identified character (a little boy). On a second or subsequent occurrence of the word group, the nominal root in the word group is coded as [nref] and the additional words receive no code in the nominal complex. Other common examples: “the bee’s hive,” or “his dog.”

+ nominal roots are the PIBU of the nominal complex and are coded according to the level of information they contain much like the verb roots discussed above. The system will recognize relatively schematic nominal roots [ngeneral] and elaborating nominal roots [nspecific].

Table 3. Common schematic [ngeneral] and elaborating [nspecific] nominals

Common schematic nouns [ngeneral] Common elaborating nouns [nspecific]
person, kid, boy Timmy (any proper name), 5 year-old
dog, frog, pet, animal(s) puppy, mutt (or specific breed name), bullfrog, tree-frog
room bedroom
jar Habitat, cage, aquarium/terrarium
chair stool, footstool
shoes boots, sneakers
ground, water grass, sand, mud, pond, puddle
tree, bushes pine, oak, bramble
animal, bird, bugs deer, gopher, groundhog, mole, beaver, rodent, owl, bees, swarm
thing, something, rock nose, head, boot, antlers, log, boulder
place, hole burrow, beehive, nest, forest, log
horns antlers
sound, noise croak, ribbit, call, howl, scream, his barking, their croaking.
other frogs/ones baby/babies, family, girlfriend, wife, lady frog, children

Plurality and countability will not impact the coding of nominal roots or pronouns (see below for pronoun coding rules)

1. The distinction between singular and plural forms of nouns (person/people, frog/frogs) and pronouns (he/they) will not impact coding. This means that plural forms of nouns are not coded as [nspecific] since they carry no more information about number than the singular forms would.

2. The distinction between mass nouns and count nouns will also not impact coding. Therefore the decision regarding [ngeneral] or [nspecific] depends on the specificity of the term, not its countability (“stuff/sugar” are both mass nouns, but “sugar” is specific while “stuff” is general. Likewise “thing/puppy” are both count nouns but “puppy” is specific)

Notice that even though plurality/countability does not impact the coding of the main nominal root, number features directly referred to by a modifier are coded. For instance if you compare “bees” with “a hundred bees” you notice that both refer to multiple bees, but that “a hundred” is more specific and contain more semantic content in the same way that “a big swarm” conveys more information than “a swarm.” “Hundred” in “a hundred bees” would carry the code [feature] (see below for more on [feature]).

+ Pronouns

Pronouns can be the PIBU in a nominal complex and are coded similarly to the referencing nominal roots, [nref], discussed above.

1. Pronouns will be coded as a special case of reference and will carry the code [pnref]. The most more common use of pronouns is to bring already shared participants, or objects back into the narrative without adding additional information about their features.

2. Pronouns can also refer to previously referenced events. This is seen in sentences like “He decided that was a bad idea” where “that” refers to a previous action (for example, knocking down a beehive). In cases such as this, “that” is simply re-referring to a previously referenced thing, in this case an event, and is coded as [pnref].

3. Some pronouns function to make reference to a general situation rather than to a particular participant, object, or event specifically referenced in the narrative previously — as in “It was scary in the forest.” Others reference the listener or story teller as in “You wouldn’t know it,” or “I think he’s hurt.” In these cases, the pronouns will be coded as [pnref]. As they are not ambiguous, no ambiguity coding is needed (see below for more on ambiguity ).

4. Pronouns occasionally refer to things that are mentioned later in an utterance in what is commonly called “backward anaphora.” An example of this that may occur in a narrative of the frog story would be the following:

He was in the room with his dog and his frog, when Timmy, the main character in this story, decided it was time to go to bed.”

In instances like this, the pronoun is still coded as [pnref] despite the reverse ordering between the nominal root and the pronoun. It is also clear that the pronoun is referring to Timmy, so no ambiguity coding is needed (see below for more on ambiguity coding).

+ Nominal Modifiers

Nominal modifiers are free standing words that convey information about the features of the PIBU in the nominal complex and will be coded simply as [feature].

Words coded as either [ngeneral],[nspecific],[nref], or [feature] can be modified by another word. An example of each type follows. The modified forms are in bold while the modifying forms are in Italics.

“The big[feature] frog [ngeneral]”

His[feature] fat[feature] bullfrog[nspecific]”

“Timmy was in his room one night. A young[feature] boy[nref] spending time with his pets.

They went into the dark scary forest. The nervous[feature] pair[nreffeature] kept close to each other.”

The very[feature] nervous[feature] pair kept close to each other.

Nominal modifiers may also modify groups of words working as a unit. So in a case where the [nref] code was applied to the second or subsequent reference to a character or object using a repeated group of words as in our “little boy” example above, a word which modifies “little boy” would carry a code of [feature]. For example the “little boy” could be given the additional feature of anger and be come “the angry little boy.” In this case the phrase would be coded “the angry[feature] little boy[nref].”

Table 4 lists some common nominal complex modifiers. Obviously this table is only a sample of the possible forms that modifiers can take. Use it as a guide to determine if a form being coded functions in a similar fashion to those in the table. As with all the tables in this manual, it is not to be seen as exhaustive.

Table 4. Common nominal modifier types [feature]

Number, Size, Shape, Color, etc… five frogs” “big deer””long log””green frog””the best one of all the baby frogs”/ “all those bees”

no sound was heard”

Placement, position, orientation, spatial features, etc… crooked branch””vast forest””front yard” / “top branch””the sound there behind the log””the one on/inside/under/beside the log”

“The hole right in front of him”

“The dog on the edge of the bed was asleep”

Texture, weight, flexibility, hardness features, etc… shiny jar” / “slimy mud””the jar that was too heavy made him fall”the soft leaves” / “the hard ground”
Class inclusion, social features, comparatives, etc… smaller frog””kid frog” / “daddy frog””terrier puppy” / “swarm of bees””bedroom window””the ugly/ pretty frog.

“the best/ funny one”

“it was one of those nothing goes right days”

“the best one of all of them”

“there were seven of them”

“The frog he found earlier in the day was in a jar.…”

Age and temporal features (older, younger), etc… baby frog””young boy””new frog””first/old pet””grown-up frog”

“the second place” / “the last place”

“the next morning” / “the last time”

Ability, attitudinal, or emotional features, etc… fast/strong deer””scary owl” / “angry bees” / “the stupid dog””he had a funny look on his face”
Relational features “the one that was on the log“”the frog without the spots“”the best one of all the baby frogs””his/the boy’s frog””It’s leg” / “the deer’s horns”

“the frog’s wife and family”

“It was on the lower ground where he fell”

“the one (that was) next to the log

“He sat on the edge of the bed”

“the log sitting right in front of him was making a funny noise”

Special Modifiers

+Connectors that relate objects/participants to each other are coded as [feature]. For example, in the sentence “The puppy and Tim went to the forest,” the connector “and” serves to relate the puppy and Tim to each other so that they can be considered a single unit acting together in the event being described. In this sense, the connector serves to modify the nouns and adds a feature to them (their relationship to each other in action). Likewise in the sentence “The puppy not Tim, ran from the bees,” the word “not” modifies the roles of the two participants in an event, this time by excluding one of them.

Connectors that simply connect two independent utterances or phrases are not coded. For example, the connector “and” is frequently used to string together utterances and would not be coded as adding specificity to either the verb or nominal complexes in the two utterances. For example in the two sentences “There was a boy and there was a dog. And they had a new frog.” the two instances of the word “and” do not specify a specific relationship between any of the actors and would not be coded as [feature]. However, in the sentence “The boy and the dog went looking for the frog” the “and” relates the two characters in action, stating that they went looking together. In this case the “and” would carry the code [feature].

This same distinction holds for connectors that relate verb complexes to each other. In cases where the connector doesn’t designate a specific relationship it remains uncoded. For example “the boy ran and the dog barked” does not provide information beyond the fact that both events occurred. Likewise in “the boy bent down and the boy looked in the hole” the “and” simple shows that both events occurred. In both examples the connector would remain uncoded. However, in “the boy ran while the dog barked” the connector “while” explains that both events occurred simultaneously and would be coded as [vsat]. Likewise in “the boy bent down and looked into the hole” the “and” relates the two events as part of a greater sequence and would carry the code [vsat].

A special relationship is that of ownership. If two nominal roots are connected via possessive form, that possessive form is coded as [feature] as in “He looked for his frog with his dog.” This would be coded as

He[pnref] looked[vmanner] for[vsat] his[feature] frog[nref] with[vsat] his[feature] dog[nref].

Possession can be ambiguous, so if it was unclear who owned the thing being referred to, the possessive form would carry an ambiguity code (see below for details on ambiguity coding).

Negation is a possible way to modify either verb complexes or nominal complexes. In verb complexes negation informs the listener that an event, state, or change of state did not occur. In nominal complexes negation informs the listener that a feature is not part of a particular object, participant, or event. In English, the forms used for negation vary and will impact how negation is coded. Freestanding satellites that convey negation in a verb complex are coded as [vsat]. An example of this would be “The frog would not[vsat] stay[vfact] in the jar.”

Freestanding forms are not used to indicate negation in English nominal complexes. Instead negation is carried by affixes to the modifier as in ” the unhappy boy” or “those antisocially inclined bees” or “the incomplete search” which are coded as [feature]. No additional coding is needed to indicate the negation.

Recognize that when negation of a feature of an object is carried by the verb complex (as in “There was not a sign of him,” or “The frog is not in the jar” or “The dog stopped being bad”) that what is being modified is the state of being verb, rather than the object/participant, event existing in that state. This means that coding of the verb complex will capture the negation which will be coded [vsat].


+ ambiguous reference in a noun, possessive, or pronoun will be recognized and coded since it impacts amount of information conveyed to the listener. If the character or object referred to with a nominal or pronoun cannot be readily identified, the code used will be [nrefambig] or [pnrefambig]. This is the procedure for both simple reference and possessives. Very little ambiguity in reference is tolerated in this coding system.

Ambiguity of nominals will frequently come from incorrect use of determiners (a/an/the/this/that/these/those). For example consider if, after already identifying a boy and a dog as being in the forest, someone said “and the boy sees a dog running from some bees.” In this case it is unclear whether this is the same dog that has been with the boy or a new dog (in fact it is most likely to be construed to mean a new character has entered the story). Likewise if a first reference to a character/object uses the article “the” this reference is ambiguous since the listener hearing “the” expects the reference to be to something previously mentioned. So, on a first reference “the boy” or “the dog” or “the jar” are ambiguous (unless the ambiguity is cleared up by other modifiers, or contextual elements — see below). Ambiguous use of reference involving nominals will be coded as [nrefambig].

An example of how contextual variables that can clarify an ambiguous nominal or pronoun is presented below.

The bees[nrefambig] chased the dog” (on a first reference to the bees)

compared to
“when he knocked down a beehive, the bees[nspecific] chased the dog”

Certain concepts such as “the ground” or “the sky” or “the woods” or “the bushes” (when in the woods) or “the window” or “the lid” (when talking about a jar) are unlikely to be confused and will tend to carry the article “the” in all instances of use without being ambiguous.

Absence of a determiner in front of a count noun can also lead to ambiguity as in “he fell in hole” since the lack of an article signals that the noun is a proper noun or mass noun (water, mud, etc…) when in fact it is not. In this example, “hole” would carry the code [nrefambig]. If the lack of a determiner is seen as an intentional use of a mechanism to make a common noun a proper name as in “they saw Frog with his family” the lack of an article does not cause ambiguity and should not be coded as [nrefambig]. In the example, “Frog” would carry the code [nref]. Notice that the general principal is that the contribution of the form to the meaning conveyed matters more than the specific form used.

Note also that a second reference to a character or object that was initially ambiguous is not necessarily ambiguous. This second reference can be coded [nref] or [pnref] IF IT IS CLEAR that it is referring to the same ambiguous object, participant, or event. An example of this follows.

The deer was there and then the deer/it carried the boy”

If this utterance were the first mention of the deer, it would be coded as

“The deer[nrefambig] was[vfact] there[vsat] and then[vsat] the deer[nref] /it[pnref] carried[vmanner] the boy[nref]”

The words “here” and “there” can be used ambiguously as pronouns as in “he looked here to see if the frog was around.” Or, “He went in there and the mouse came out.” In these cases, if it is unclear where “here” or “there” refers to, these words would be coded as [pnrefambig].

Possessives will carry an ambiguity code when it is unclear who the owner in the relationship is. For instance, if a child said “They saw the deer then he got on his head” without appropriate contextual support, then it may be unclear whose head is being talked about. In this case the sentence would be coded as follows:

They[pnref] saw[vmanner] the deer[nref] then[vsat] he[pnrefambig] got[vfact] on[vsat] his[pnrefambig] head[nspecific].


Recognize that there are well-formed uses of reference such as “It was scary in the forest” or “it seemed hopeless” that do not diminish the semantic content or increase the semantic content significantly despite some ambiguity about the exact referent of the pronoun. These types of reference will not be coded as ambiguous in this system. These cases are taken care of with the code [pnref] as discussed above.

+Other factors +Articles/determiners will not be coded in this system, as they are obligatory when used correctly in English constructions (See discussion above regarding ambiguous reference). The only exception to this rule is when they are part of a complex construction (see below). Determiners include “a,” “an,” “the” and (sometimes) “this/that/these/those.”

Contractions will be thought of as split into their constituents for coding. This means that a contraction such as “he’s” is coded as if it were “he is” or “they’re” is coded as if it were “they are.” Sometimes in spoken language contractions will be built using nouns as well as pronouns. This means that a contraction such as “the boy’s in the tree” will be coded as if it said, “the boy is in the tree.” To code the utterances, place codes in sequential order after the contraction.

“the boy’s[nref][vfact] in[vsat] the tree[nref]”)
“the deer’s[nrefambig][vfact] in[vsat] there[pnrefambig].”

Another example would be “he couldn’t keep running” the “could” in the contraction is a verb satellite carrying modal information so the example would be coded as “couldn’t[vsat][vsat] keep[vsat] running[vmanner].”

Another example would be “the deer’s running” where the “is” is a simple tense marker. In this case code the utterance as “the deer’s[nref][vfact] running[vmanner].”

Complex constructions will be coded as collections of individual words despite the fact that they work as a unit. For example, in the sentence “No matter how hard he looked, he couldn’t find his frog in his room,” the construction “no matter how hard” works as a single unit to indicate the degree or intensity of looking occurring in the event. For coding purposes each word in this construction would be coded as a verb satellite ([vsat]) despite the fact that the construction works as a multiword unit. This is true whether the constructions are broken up across the sentence or not. Some more examples include “If he would have left the jar on tighter, then the frog wouldn’t have gotten away,” or “The more he looked, the less sure he was.”

Complex constructions can also help to modify the meaning of the nominal complex as in “He put on his favorite looking for frog shoes.” In these cases, each word would carry a code of [feature] despite the fact that the construction works as a unit. Notice that the construction in the example could be written as “looking-for-frog shoes.” Identify constructions that modify the nominal complex and could be partially or completely hyphenated, then code the words individually as [feature]. Other examples would include “he heard a frog like sound” or “he was in a holier than thou mood” or “it was one of those nothing goes right days” or “he got caught in the horns that looked like branches on the deer’s head” or “The hole right in front of him had a gopher in it.”

Notice that complex constructions can come in the form of dependent clauses, or prepositional phrases. For example, nominals are often specified with reference to their participation in a previous event such as “They were looking for the frog that used to be in the jar.” When a clause or phrase is used to specify the meaning of a verb complex or noun complex, then all words in that phrase or clause will carry the appropriate modifier code. In this example the sentence would be coded

They[pnref] were looking[vmanner] for[vsat] the frog[ngeneral] that[feature] used[feature] to[feature] be[feature] in[feature] the[feature] jar[feature].”

A more complicated example,

He[pnref] grabbed[vmanner] onto[vsat] the thing[ngeneral] that[feature] looked[feature] like[feature] branches[feature] but[feature] was[feature] actually[feature] a[feature] horn[feature]”

Compare this to “He grabbed onto the thing that looked like branches, but it was actually a horn.” In this case the re-referencing of “the thing” with the pronoun “it” breaks up this conceptual unit. In this case the utterance would be coded as

“…but[feature] it[pnref] was[vfact] actually[vsat] a horn[nspecific].”

An example of a clause modifying the verb complex is found in a sentence like “The dog ran because the bees were chasing him.” In this case the sentence would be coded

“The dog[nref] ran[vmanner] because[vsat] the[vsat] bees[vsat] were[vsat] chasing[vsat] him[vsat]”

This is done because the clause is helping to specify the meaning of the verb by indicating a cause for the action. Compare this to the sentence “The dog ran and the bees were chasing him.” This second sentence would be coded as a series of two events as indicated below.

“The dog[nref] ran[vmanner] and the bees[nref] were chasing[vmanner] him[pnref].”

In this case the same clause is being used differently. It is not stating the cause of the running. Instead it is merely indicating another co-occurring sub-event in the scene. Another example would be “He went where the boy was going” coded as

he[pnref] went[vfact] where[vsat] the[vsat] boy[vsat] was[vsat] going[vsat].”

Watch for constructions starting with forms like “when, where, because, cuz, after, before, etc…” that will tend to modify verb complexes, and for constructions starting with forms such as “which, that, who” that will tend to modify nominal complexes. Also recognize that sometimes these forms are left out as in “He discovered the frog[vsat] had[vsat] escaped[vsat].” Also watch for ambiguous reference within a complex construction as in “He[pnrefambig] discovered[vmanner] he[pnrefambig] had[vsat] escaped[vsat]” where “he” is ambiguous. AMBIGUITY CODING TAKES PRECEDENCE IN THESE CASES AND THE WORD IS CODED ACCORDINGLY.

Poorly Formed Utterances

Poorly formed utterances as in “Then a long ways down to get to the water then they fell” can complicate coding decisions. In these cases, code elements in terms of their role despite the difficulty in organization among the elements. Any element that cannot be assigned a clear role should be left uncoded. This example might be coded as follows.

“Then a long[vsat] ways[vsat] down[vsat] to[vsat] get[vfact] to[vsat] the water[nref] then[vsat] they[pnrefambig] fell[vmanner].”

Part of the organizational trouble in this utterance comes from the repetition and poor placement of “then” (and /or the lack of “it was”). Since it is not clear how the initial “then” relates to the other elements in the sentence it carries no code. The second “then” clearly conveys temporal information relating the two happenings in the utterance (the getting and the falling) so it would carry a code of [vsat] to indicate this contribution in clarifying (admittedly in an awkward fashion) the relationship between the verb complexes.

Watch carefully for phrases or clauses that rename something referred to initially with a general noun as in “He saw a boy, the frog that he had before, and a girl frog” which seems to be a revision of a simpler construction such as “a boy frog.” In this example, “the frog that he had before” serves to modify “boy” as a unit, and each word would carry the code [feature]:

He[pnref] saw[vmanner] a boy[nref] the[feature] frog[feature] that[feature] he[feature] had[feature] before[feature] and a girl[feature] frog[ngeneral].

Notice that this is an instance of contextual information clearing up an initially ambiguous form (boy).

Character Speech will be coded as a special case of a complex satellite construction. This means that all the words attributed to a character will carry the code [vsat] as they are in essence modifying or specifying the meaning of the verb “said” or “told” or “yelled” just as other satellites might. For example, in a sentence such as “The boy told his dog not to do that again,” “not to do that again” modifies the verb “told” and provides the listener with information not available in the sentences “The boy told his dog” or “The boy told his dog something.” Likewise, in a sentence such as “The boy told the dog ‘You better not do that again'” the complex construction “you better not do that again” modifies the verb “told” giving the listener more information than the sentence without the construction. In these cases each word in the character speech will carry the code [vsat]. This holds true for cases such as “He was thinking it look scary in there,” or “He thought, I don’t know where else to look


This is also true of sound effects in the story as in “they went “SPLASH!” Sound effects are helping to specify the verb, in this case “went” and are counted as [vsat]. Other example might include “it was all “EEEE!” or “Sploop! They went right into the pond.”

Feature or Satellite? The distinction between nominal feature and verb satellite is not always easily made. For example consider the following sentences.

1. The log was sitting right in front of him making a funny noise.

2. There was a log right in front of him making a funny noise.

3. The log right in front of him was making a funny noise.

4. The log sitting right in front of him was making a funny noise.

Is the word “right” in each of these sentences a nominal feature or a verb satellite? Notice that the construction “right in front of him” does not vary in form from sentence to sentence, but the coding for each sentence will be different. In sentences 1 and 2, the construction is helping to elaborate what is occurring (the log was existing or sitting) while in sentences 3 and 4 the construction is helping to specify which log is making a funny noise. So “right” is coded [vsat] in sentences 1 and 2, but as [feature] in sentences 3 and 4. Let’s complicate the picture with some more sentences.

5. Sitting there right in front of him was a log making funny noises.

6. The log sitting there right in front of him was making funny noises.

This time around, the word “right” in sentence 5 carries the code [vsat] since it specifies where the log is sitting, while in 6 it carries the code [feature] since it helps specify which log is making the noise. It is not the form of the word that matters but its function in the sentence that determines which code it carries. Also note the word “sitting” in sentences 5 and 6. In sentence 5, the words “making” and “sitting” both elaborate what the log is doing, while in 6, only “making” elaborates what the log is doing, while “sitting” is indicating which log is doing it. So in sentence 5 both “sitting” and “making” are PIBU’s and carry the code [vmanner] or [vfact] respectively. While in sentence 6 only “making” is a PIBU carrying the code [vfact]. (see the section below with training examples for exact coding of each of these utterances.)

Extra-Narrative Comments not related to the story telling will not be included in the analysis. For example, utterances such as “This picture’s kind of confusing” or “I don’t know what to say” or “What are you doing that for?” or “What time is it?” or “I haven’t eaten lunch yet,” or “…because I can’t think of anything else” or “What’re you doing?” or “Once upon a time” or “The end” or “one, two, three, four, five, six…” do not relate to the events in the story being related, and will not be included in the analysis.

Comments that have a direct impact on the information conveyed in the story will be coded as part of the story. For example “His name was Jill, no, no, I don’t know.” In this case the extra-narrative comment (“no, no”) has a direct impact on the information conveyed to the listener and will be coded (conveying that the character’s name is not known). None of the references are ambiguous, so the sentence would be coded as

His[feature] name[nspecific] was[vfact] Jill[nspecific] no[vsat], no[vsat], I don’t know.”

Other examples include “I think I’ll call him Froggie” or ” I wasn’t sure before, but I think they’re bees.” These types of utterances are sometimes called hedges, and are part of the story telling process. The distinction is that the hedge is about a story character, object or event, and not about the story-teller (as in “I don’t know”). Hedges convey the degree of certainty with which the information should be taken, and are therefore part of the story telling process. For instance “I think I’ll call him Froggie” informs the listener that this is just one of many frog names that could have been used, rather than the frogs “real” name. Likewise, “I think they’re bees” informs the listener that “they” are bees or something very similar to bees. Pragmatically, this gives the parameters the listener should be using in framing an understanding of events. The pronoun “I” in these cases is unambiguous and would be coded as [pnref].

Mazes and the words that make them up are not coded in this system. Mazes are defined as false starts and repetitions or revisions in an utterance. In the transcript these will be put inside parenthesis and excluded from analysis.


Croft, W. 2001. Radical construction grammar: syntactic theory in typological perspective. Oxford: Oxford University Press.

Lakoff, G. 1987. Women, fire and dangerous things: what categories reveal about the mind. Chicago: The University of Chicago Press.

Langacker, R.W. 1991. Concept, image and symbol: the cognitive basis of grammar. Berlin: Mouton De Gruyter.

Mayer, M. 1969. Frog, where are you? New York: Dial Press.

Talmy, L. 2000. Toward a cognitive semantics, volumes I and II. Cambridge, MA: MIT Press.

Thordardottir, E.T., & Ellis Weismer, S. 2001. High-frequency verbs and verb diversity in the spontaneous speech of school-age children with specific language impairment. International Journal of Language & Communication Disorders, 36 (2) 221-244.

Additional Tables:


Table 6: Summary of Codes

Table 1a. Recognizing elaborated verbs [vmanner]

Determine if the verb root elaborates–

the manner in which something happens/ happened/ will happen.

He walked/ran/strolled down the road” (compare to went)

His frog escaped/ tip toed from the jar. (compare to got out)

He jumped on the rock. (compare to goton)

He tripped/fell/stumbled. (compare to “He is on the ground.”)

He grabbed/climbed a branch. (compare to got/ got on)

He searched/looked for his frog. (compare to “He went for his frog.”)

He thought about it. (compare to “He is there.”)

He lost his frog. (compare to does not have)

He yelled/whispered/screamed. (compare to told)

Indicate the manner in which something exists/ existed/ is/will be in a state.

He lived with a boy and his dog. (compare to is)

They slept while the frog escaped. (compare to were there)

The boy was worrying about his frog. (compare to didn’t have)

He seems angry. (compare to “He is angry)

Indicate the manner in which something changes/ changed/ will change state.

Then the jar shattered. (compare to broke)

The dog was nearly killed by the fall. (compare to died)

He woke up. (compare to got up)

The hive was crushed. (compare to was in pieces)

If so they are coded as [vmanner].

Table 6 summary of codes

additional tables

Verb Complex Nominal Complex
[vfact] *use for schematic verb roots [ngeneral], [nspecific]*use on 1st reference to object/participant
[vmanner] *use for elaborating verb roots [nref]  [pnref]*use for nth reference to object/participant*[pnref] also used in reference to general situation like “It was scary in the forest.”
[vsat] *use for verb satellites that elaborate the meaning of the verb complex [feature]*use for elements that elaborate the meaning of the nominal complex
[nrefambig], [pnrefambig] *use for ambiguous noun/pronoun/ possessive’s

Note on “the window” and ambiguity.

The window is a marginal case similar to “the floor” or “the door” or “the sky.” It is very infrequent for a child to introduce the window with a clearly elaborated form such as “his window” or “the bedroom window.” It is also infrequent for the child to take advantage of the a/the distinction when introducing the window. For this reason, the form “the window” will be treated the same as “the floor” or “the ground” or “the sky” and will be coded as [nref] or [ngeneral].

There[vsat] was[vfact] no[feature] sign[ngeneral] of[feature] frog[nref]. They[pnref] looked[vmanner] again[vsat]. No[feature] sign[ngeneral] of[feature] frog[nref]. NOTE: “no” in “no sign” tells how many signs there are which makes it a [feature]. Each time they look a different sign doesn’t appear, so the second “sign” is a [ngeneral].

45a. They[pnref] found[vmanner] a whole[feature] family[feature] of[feature] frogs[ngeneral]. NOTE that “whole family of” works just like “whole bunch of” or “some” or  “a lot of” to tell how many frogs there are (with some added relational features). So “family” gets a code of [feature] since it tells about features of the frogs.

NOTE on the bottle-jar relationship in reference. Bottle and Jar are both used to refer to the glass container that the frog is held in. If the container is first referenced as one, and then as the other, it is often ambiguous as to whether this is the same container or a different one. In the case where is it unclear, these get coded as [nrefambig].  Sometimes the child tries to disambiguate by adding features to this ambiguous reference as in “the bottle that the frog was in” to refer to what was previously called a “jar.” This may not be sufficient. If it is still possible that we are talking about a jar and a bottle (when there is really only one in the story) then it gets coded as [nrefambig].  If it is clear they are the same container (rarely is it), code as [nref].

+ There constructions
The word there has two basic meanings in English: the deictic and the existential. The deictic is thought to be the most basic and is used to point to place (It’s right there); pointing often accompanies deictic there in face to face conversation. The existential there is an abstracted extension of the deictic and is used to point to the existence of something. Here are some examples of each.

Deictic: The frog was there behind the log.
There the boy is. He is with his dog in the puddle.
He looked behind the log, but it wasn’t there.
There it is, behind the log, and the boy found it.
Existential: There was a boy who lost his frog.
There’s a gopher in the hole who bit him.
There’s really a deer behind the rock, but he doesn’t know.

However, in our coding system this distinction is not really the one that is most important.  This coding system pays attention to the use of language to elaborate concepts (or make them schematic).  So there can be a problematic construction to code.  As discussed above, there is most often a [vsat] since it locates the place of action (deictic) or existence (existential).  Sometimes decitic there, however, is used to refer to a specific location very schematically as in the third (bolded) example above. In these types of cases, there works more like a pronoun reference than a verb satellite and will carry the code [pnref]. In cases where this type of usage is ambiguous, there will carry the code [pnrefambig]. An example of an ambiguous pronoun there would be:

He looked but it wasn’t there[pnrefambig].

If it is the case that there is not enough contextual support to specify where he was looking.

The best way to identify these cases is to replace the “there” (or “here”) with the place you think it might be referring to. For instance, in the bolded example above, since you can find “behind the log” and place it in the sentence to get

He looked behind the log, but it wasn’t behind the log.

You can code there as [pnref].  However, in our ambiguous example, “behind the log” is not available and the listener does not know where “he looked” so there is coded [pnrefambig].

This only works with deictic there. Existential there essentially locates the existence of the thing being referred to in the mental space of the narrative (Lakoff, 1987). In other words, existential there places the existence of the character or object in the story.  If you can add the words “in the story” into the sentence and still have it make sense, then there is probably existential and carries the code [vsat]

In the story was a boy who lost his frog.
In the story is a gopher in the hole who bit him.
In the story is really a deer behind the rock, but he doesn’t know.

+ After as a verb.

A common constuctions used by children when describing the bees chasing the dog in the story involves the word “after.”

The bees were after the dog.
They’re after him.
The owl’s still after him.


They got after the dog.
The bees got mad and went after the dog.

This is a tricky construction to code because “after” is used to mean “chasing,” a manner verb. Therefore “after” which is usually a [vsat] gets coded as [vmanner] when it means “chasing.” And like “chasing” it requires a tense which is carried by the auxilary verbs “is, was, were…” which are coded vfact.

From the examples above.

The bees[nref] were[vfact] after[vmanner] the dog[nref].
They‘re[pnref][vfact] after[vmanner] him[pnref].
The owl’s[nref] still[vsat] after[vmanner] him[pnref].
The bees[nref] got[vfact] mad[vsat] and went[vfact] after[vsat] the dog[nref].


They[pnref] went[vfact] after[vsat] the dog[nref].
The bees[nref] got[vfact] mad[vsat] and[vsat] flew[vmanner] after[vsat] him[pnref].

In these cases, “after” is back to its normal role as a [vsat] and is elaborating schematic verbs “flew” and “went.”
Notice that when it is a [vsat] “after” does not require the auxilary verbs to indicate tense.

Coding Protocols to reduce errors

When coding in SALT, use the following steps to reduce common errors.

STEP ONE: Read through the entire story and identify any poorly formed utterances. Write these down on a piece of paper, and edit them, crossing out un-necessary words, so that the best possible interpretation of the utterance can be used to guide you in coding when you come to these utterances. Remember, do not code any words that you can not clearly identify a role for in the utterance (these would be the words you crossed out when editing).

STEP TWO: Using SALT’s search function, search for words that commonly indicate the existence of a relative clause such as “where, when, because, ’cause, cuz, after, before” and “that, which, who, what.” This should help you identify some of the complex constructions that will be coded as a unit, with every word getting a code of either [feature] (for “that, which, who, what” clauses) or [vsat] (for “where, when, because, ’cause, cuz, after, before” clauses). Remember, not every instance of these words will be part of a complex construction or relative clause. USE THE EXAMPLES PROVIDED TO HELP YOU DETERMINE HOW TO CODE THESE RELATIVE CLAUSES.

STEP THREE: Using SALT’s search function, search for “and.” Code instances of “and” that fall between two things (i.e. the boy and the dog”) as [feature], code instances of “and” that fall between two actions (i.e. They looked and looked,” or “he ran and fell”) as [vsat].  Instances of “And” that connect two sentences, clauses or utterences (i.e. “And the boy was looking.” or “The boy ran and the dog was getting chased”) do not carry a code.

STEP FOUR: Using SALT’s search function, search for all the auxilary verbs (is, are, was, were, will be) and the contractions negating them (isn’t, aren’t, weren’t, wasn’t) and UNLESS THEY WERE ALREADY CODED IN STEP TWO,  code them: is[vfact], are[vfact], was[vfact], were[vfact], will[vfact] be[vfact]; isn’t[vfact][vsat], aren’t[vfact][vsat], wasn’t[vfact][vsat], weren’t[vfact][vsat], won’t[vfact][vsat] be[vfact].

STEP FIVE: Code the story start to finish.

STEP SIX: Use SALT’s analyze function to generate a CODE TABLE EXPANDED BY WORDS. Use this to double check your coding.
a) Make sure all the codes that show up are legitimate codes (usually extra codes are due to spelling errors, e.g.[vast])
b) Make sure all [nspecific] words are specific nouns, and that all [ngeneral] words and general nouns.
c) Make sure only pronoun’s show up in [pnref] or [pnrefambig]. Watch for possessive pronouns which should be coded as either [feature] or [pnrefambig].
c) Generally assess the coding of specific words. Check questions in context.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s