Can the smallest possible set of words ("semantic primitives") be selected, such that those words alone can be used to define all the remaining words in a language?
Most broad-coverage dictionaries (such as the American Heritage Dictionary) make no attempt to use semantic primitives, so all their word definitions are ultimately circular.
The Longman Dictionary of Contemporary English, however, uses a 2000-word defining vocabulary to define all the remaining 54,000 words in the dictionary (without relying on idioms or circular definitions). But the defining vocabulary is not as small as possible: many of the words (such as "mother," "daughter," "sister") can be defined using other words ("female," "parent", etc.) from the defining vocabulary.
Attempts to select smaller sets of primitives have either been incomplete or non-compositional. Jackendoff's Lexical Conceptual Structure and other systems of primitives (Wilks, Schank, etc.) are incomplete: they cannot define all words in a broad-coverage dictionary. Ogden's 850-word Basic English vocabulary can be used to define a broader range of words, but it does so by being non-compositional: it uses idioms (such as "put up with" meaning "tolerate") and words with multiple meanings that depend on context.
Anna Wierzbicka and Cliff Goddard's system of Natural Semantic Metalanguage Primitives may also be incomplete, but it is based on cross-linguistic studies of language universals, and has been used to create clear definitions for hundreds of words. (See "Semantics: Primes and Universals," Wierzbicka, 1996. Also see The Natural Semantics Metalanguage Homepage.) This set of primitives is listed below: each primitive and its variations (in bold caps), followed by example sentences (in italics), and the primitive's argument slots (X1, X2, X3, X4).
I am interested in bridging the gap between a small but incomplete set of primitives (like the Natural Semantic Metalanguage Primitives) and the broad-coverage Longman defining vocabulary. What small number of additional semantic primitives are needed to define the entire Longman vocabulary (which in turn can define all other words)?
If you share an interest in broad-coverage compositional semantic primitives, please send me email. I would enjoy discussing ideas with you.
Thanks, David Bullock
[email protected]
Substantives:
I.
I think about you.
X1 = One's self.
YOU, THOU.
You think about me.
X1 = The person being addressed.
SOMEONE, WHO, BEING.
Someone did something bad to this person. Who is this?
X1 = A person or being.
SOMETHING, THING, WHAT.
Something happened to this thing. What did you say?
X1 = A thing.
PEOPLE.
These people moved.
X1 = Humans.
BODY.
Eyes are parts of your body.
[Added, 2002]
X1 = A person or animal.
X2 = The physical body of X1.
Determiners and Quantifiers:
THIS.
This happened to me.
X1 = Something indicated by the speaker.
THE_SAME, THE_SAME_AS.
It was not the same fish, but it was the same kind of fish.
X1 = Something that is the same thing as X2.
X2 = Something that is the same thing as X1.
OTHER, ANOTHER.
I saw it, and two other people saw it.
X1 = Something that is not the same as X2.
X2 = Something that is not the same as X1.
ONE, ONE_OF.
I have one daughter. One of these people moved.
X1 = A set containing exactly one element.
X2 = The superset of which X1 is a subset.
TWO.
I see two things.
X1 = A set containing exactly two elements.
MANY, MUCH.
Many people think this. I didn't say much.
X1 = Something that is in a large quantity (subjectively).
ALL.
All people want this.
X1 = Symbol.
X2 = Definition of a set.
X3 = Statement that is true about each element X1 in set X2.
SOME, SOME_OF.
Some people know this.
X1 = Symbol.
X2 = Definition of a set.
X3 = Statement that is true about at least one element X1 in set X2.
MORE.
I want to see more.
X1 = Something that is in a larger quantity than X2.
X2 = Something that is in a smaller quantity than X1.
Mental Predicates:
THINK, THINK_ABOUT.
I think: I don't want this. I think about you.
X1 = Person who is thinking.
X2 = Statement that X1 believes to be true.
X3 = Topic about which X1 is thinking.
KNOW, KNOW_ABOUT.
I know: I can't do anything. I know about this place.
X1 = Person who knows something.
X2 = Statement that X1 knows to be inherently true.
X3 = Topic about which X1 knows something.
WANT.
I want to say something now.
X1 = Person who wants something.
X2 = Statement describing what X1 wants.
FEEL.
I feel something bad.
X1 = Person who is perceiving or experiencing.
X2 = Statement describing what X1 feels.
SEE.
I see many people.
X1 = Person who is seeing.
X2 = What X1 is seeing.
HEAR.
I hear something. I hear music.
X1 = Person who is hearing.
X2 = What X1 is hearing.
Speech and Words:
SAY, SAY_TO, SAY_ABOUT.
I said one word. He said something to Mary. I want to say something about these people. I said: I want this.
X1 = Person who is saying something.
X2 = What X1 is saying using vocal or other signs.
X3 = The person to whom X1 is saying X2.
X4 = Topic about which X1 is saying something.
WORD.
You said some bad words to this person.
X1 = A word.
TRUE.
I know that this is not true.
[Added, 2002]
X1 = Statement that is true.
Actions, Events, Movement, Existence, Possession, Life:
DO, DO_TO, DO_WITH.
You did something bad. You did something bad to me. I did this with a knife.
X1 = Agent performing some action.
X2 = Action that X1 is performing.
X3 = Patient to whom X1 is performing action X2.
X4 = Instrument with which X1 is performing action X2.
HAPPEN, HAPPEN_TO.
Something bad happened in this place. Something happened to me.
X1 = Action that is happening.
X2 = Patient to whom action X1 is happening.
MOVE.
Something moved here.
X1 = Something that is moving.
THERE_IS, THERE_ARE, BE.
There are three kinds of fish here.
X1 = Something that exists.
HAVE, BELONG_TO.
I have a dog. This cup belongs to me.
[Added, 2002]
X1 = Possessor.
X2 = Possession.
LIVE, ALIVE, LIVING_THINGS.
These people lived for a long time.
X1 = Something that is alive.
DIE.
Someone died at this time.
[Added, 2002]
X1 = Something that died.
Evaluators and Descriptors:
GOOD, GOOD_FOR.
This is good. This is good for you.
X1 = Something that is objectively good.
X2 = Someone for whom X1 is beneficial.
BAD, BAD_FOR.
You are a bad person. This is bad for me.
X1 = Something that is objectively bad.
X2 = Someone for whom X1 is harmful.
BIG.
These are big people.
X1 = Something that is big (subjectively, relative to similar things).
SMALL.
This is a small dog.
X1 = Something that is small (subjectively, relative to similar things).
Time:
WHEN, TIME, AT_A_TIME.
When I did these things, I felt something bad. At that time, I heard something. When did you do this?
X1 = A time.
X2 = Statement describing something happening at time X1.
BEFORE.
This happened before I saw you.
X1 = Statement or time occurring before X2.
X2 = Statement or time occurring after X1.
AFTER.
I was born after you were born.
X1 = Statement or time occurring after X2.
X2 = Statement or time occurring before X1.
A_LONG_TIME.
It happened a long time before now.
X1 = A long period of time (subjectively, relative to comparable time periods).
X2 = Statement describing something happening during time period X1.
A_SHORT_TIME.
I did this for a short time.
X1 = A short period of time (subjectively, relative to comparable time periods).
X2 = Statement describing something happening during time period X1.
NOW.
This person is moving now.
X1 = The current time from the speaker's perspective.
X2 = Statement describing something happening during time period X1.
FOR_SOME_TIME.
I did this for some time.
[Added, 2002]
X1 = A period of time with a duration longer than a moment.
X2 = Statement describing something happening during time period X1.
MOMENT.
I saw her for a moment. A moment later he died.
[Tentative, 2002]
X1 = An instant in time.
X2 = Statement describing something happening at time X1.
Space:
WHERE, PLACE, IN_A_PLACE, BE_IN_A_PLACE, BE_SOMEWHERE.
I know where it is. It happened in another place. Where are you?
X1 = A place.
X2 = Statement describing something happening in place X1.
UNDER, BELOW.
This thing is under this other thing.
X1 = Place or thing that is under X2.
X2 = Place or thing that is above X1.
ABOVE.
The head is above all other parts of a person's body.
X1 = Place or thing that is above X2.
X2 = Place or thing that is under X1.
FAR, FAR_FROM.
This thing is very far from me.
X1 = A place or thing that is far from X2 (subjectively, relative to comparable distances).
X2 = Another place or thing.
NEAR, NEAR_TO, CLOSE_TO.
This person is near to the other person.
X1 = A place or thing that is close to X2 (subjectively, relative to comparable distances).
X2 = Another place or thing.
SIDE, ON_ONE_SIDE_OF, ON_THIS_SIDE, ON_WHAT_SIDE.
He is on this side of me, and she is on the other side.
X1 = Something that is on side X3 of X2.
X2 = A place or thing.
X3 = The side of X2 where X1 is located.
X4 = The other side of X2.
INSIDE.
I was inside the house.
X1 = Something that is inside X2.
X2 = A container.
HERE.
I am here.
X1 = The current place from the speaker's perspective.
X2 = Statement describing something happening in place X1.
TOUCHING.
Something is touching this part of my body.
[Tentative, 2002]
X1 = Something that is touching X2.
X2 = Something that is touching X1.
Partonomy, Taxonomy, Similarity:
PART, PART_OF.
A foot is part of someone's leg.
X1 = Something that is a part of X2.
X2 = Something that has parts.
KIND, KIND_OF.
This tree does not look like that other tree, but they are the same kind of tree.
X1 = Something that is a member of category X2.
X2 = A category.
LIKE, AS.
I am not like other people.
X1 = Statement or thing that is similar to X2.
X2 = Statement or thing that is similar to X1.
Metapredicates, Clause Operators and Linkers:
NOT.
This is not in the same place.
X1 = Statement that is not true.
CAN.
This thing can move.
X1 = Statement that is possible.
VERY.
This is very good.
X1 = Statement that is intensely true (subjectively).
IF.
If you do this, people will say something bad about you.
X1 = Statement.
X2 = Statement that is true if X1 is true.
BECAUSE.
The dog died because the man hit it on the head.
X1 = Statement describing an effect of X2.
X2 = Statement describing a cause of X1.
IF_WOULD.
If I were you, I would not do it.
[Removed, 2002]
X1 = Statement that is not true.
X2 = Statement imagining what would be true if X1 were true.
MAYBE.
Maybe all these people were in the same place at this time.
X1 = Statement that might be true, but is not known to be true.