The Semantic Tangle
By Jessica Lake
I have been thinking about semantic knowledge.
Habitually.
I use “knowledge” here because when people speak of sweet memories, I take them to mean something they can revisit.
There are no memories here. This is the Land of Knowledge.
After months of musing, my conception of what’s up with knowledge is beginning to mature.
So I would like to share it with you.
What immediately follows is a list of basics, each building on the one before it. This will give you the footing.
Knowledge applies lossy compression to heterogeneous episodes. It rolls them up into statistical regularities. Life without the living.
These regularities are stored and organized within semantic knowledge for flexible associative access.
It seems clear that knowledge takes several forms.
Constraint — A declaration that prohibits a feature. Some constraints are expressed in words, others in abstraction.
Lexicon — This is the treasure trove of spoken language. A searchable catalogue of spoken words and their uses.
Syntax template — Syntax suggests itself as a series of templates expectant of completion by thought. Think Mad Libs.
Fact — A brief statement of something of merit. It could be heard once or many times.
Regularity — A compression of recurrent experience. Childhood summers, school days, or time spent with your mother.
Rule — A payload fired by a pattern trigger. Behavior that responds to current circumstances.
Model — A representation of statistical regularity expressed as a transform for prediction.
Semantic Representation
Some things are hidden from me within knowledge. Others are obvious. I can see words and meanings. But abstractions elude me.
My vision of semantic knowledge does not express itself as a classical DAG or a high-dimensional space. My nature is to prune relentlessly.
Imagine knowledge as a diffuse tangle of millions of colored strings in a void. Each string a unique constraint. You search for objects by what they are not.
You’re trying to find a word. No, that’s not right. That’s too rigid. No. Something softer. But not too soft.
Constraint search is natural.
Do not sweat the void. It is amorphous and sufficient to hold the tangle. It permits imagination.
Axiom — Any known can be isolated at the intersection of its constraints.
The constraint intersection is recorded with a tag embossed with a word or phrase. A tag gives the known an identity so it may be referenced. The constraints involved form a constraint-set.
DAG — Who doesn’t love a directed acyclic graph? You traverse its directed edges, passing through nodes. The pathway is a sequence of decisions very much like Candy Land.
Node — A node is a place of rest. It may afford properties. If you identify a node by the path you took to reach it, changing the path changes its address. Add or remove a node along the path.
What happens to the addresses of successive nodes?
The Tangle — A tag’s identity is declarative. You light up its constraint-set. You apply the constraints in any order, or all at once. No traversal is required.
All that distinguishes one tag semantically from another is the particular set of constraints that intersects there.
Definition — The Tangle isolates a tag by its constraint-set, written {C}. It requires no coordinates, ordered tuples, or traversal. The set is unordered.
{C} Constraint-Set
The complete set of constraints defining a tag is called its {C}.
A {C} can therefore stand alone as a specification of a tag. It says what the tag is entirely in terms of the constraints that define it.
A tag has no semantic identity independent of its {C}. It does, however, have an identity for reference.
This reverses the usual intuition of a semantic network. In a network, an identifiable node exists and relationships connect that node to other nodes.
Here there are no semantic nodes.
Identity arises from the intersection of the constraints that define a tag.
Constraint Convergence
A single constraint may apply to many tags.
Applying that constraint therefore isolates a population of tags sharing the same prohibition.
A second constraint reduces that population to those tags sharing both prohibitions.
Additional constraints narrow the surviving population whenever they rule out some of the remaining tags. The same constraints leave the same survivors whether applied one by one or all at once.
Constraint Convergence can therefore terminate at different levels of specificity.
If the applied constraints distinguish a single tag, the convergence identifies that tag.
If several tags survive, the result is a group of tags sharing the applied prohibitions.
Thus, the same mechanism can identify either an individual regularity or a group of related regularities.
One constraint → a population of tags Additional discriminating constraints → a smaller population Sufficient constraints → a single tag
A search need only apply enough constraints to isolate a tag. The tag’s identity still comes from its full {C}.
Learning
The architecture is inherently extensible.
Observations may reveal regularities that cannot be adequately distinguished using the constraints already present in semantic knowledge.
When this occurs, a new constraint can be created.
Adding it makes a new distinction available. To distinguish two existing tags, the model still has to establish which of them satisfies the new constraint.
That constraint can then be used in new combinations, making distinctions that could not previously be made.
Existing semantic knowledge does not necessarily have to be reconstructed.
Instead, learning can progressively increase the resolution of the existing representation by adding constraints.
New constraints → new intersections → finer discrimination
This permits semantic knowledge to become increasingly differentiated as experience accumulates.
Realms
For tractability, the complete semantic representation can be partitioned into realms.
A realm corresponds to a particular aspect of reality.
Realms serve a practical purpose. Partitioning reduces collisions between unrelated operations and permits asynchronous reads and writes to different portions of semantic knowledge.
Categories emerge from shared constraints.
Realms partition the machinery that stores and processes those constraints.
In Closing
The model can be reduced to several propositions.
Cognition extracts statistical regularities from reality.
Semantic knowledge stores those regularities.
A regularity may grant or deny attributes.
Prohibitions are represented as constraints. Constraint Convergence operates on those constraints.
The Tangle represents semantic knowledge through intersections of constraints.
Constraints can be applied in any order.
A tag represents an intersection of constraints and gives it an identity for reference.
The complete set of constraints defining a tag constitutes its {C}.
Applying constraints isolates tags sharing those prohibitions.
A search may return a group of tags or isolate a single tag. It can isolate a tag without using the full {C}.
New knowledge can introduce new constraints, increasing semantic resolution without requiring wholesale reconstruction of existing knowledge.
Realms partition the machinery for concurrency and collision reduction. Categories emerge from shared constraints.
Semantic identity emerges from the cumulative intersection of constraints.