Toward Human-Level Cognitive Adequacy
Our long-range aim is to design and implement common sense in a
computer. Click here
for details.
If you would like to learn more about Active Logic, we suggest
you start with one of our primers.
There are several examples of active logics in our papers. We present here a couple of simple ones.
| \(t\): | \(Now(t)\) |
| \(t+1\): | \(Now(t+1)\) |
is a rule that says: if at the current step, \(Now\) has the value \(t\), then, at the next step, let \(Now\) have the value \(t+1\). This enables the active logic to keep track of step numbers and therefore of time.
This is a basic rule and is included in all active logics.
Another example is the contradiction rule:
| \(t\): | \(P, \neg P\) |
| \(t+1\): | \(contra(P,\neg P)\) |
If at a step, we have both \(P\) and \(\neg P\) present in the database, at the next step, we add \(contra(P,\neg P)\) to the database to indicate the contradiction. There will be other rules that will cause the consequences of \(P\) and \(\neg P\) not to be derived in later steps, and rules that will attempt to resolve the contradiction and reinstate either \(P\) or \(\neg P\) to the database at a later time.
We can also have modus ponens:
| \(t\): | \(P, P\implies Q\) |
| \(t+1\): | \(Q\) |
This says: if at time \(t\), the database contains \(P\) and \(P\implies Q\), then in the next time step, conclude \(Q\).
Note that if the database contains \(P\), \(P\implies Q\), and \(Q\implies R\), we do not get \(R\) immediately, but only after 2 steps. First, we use \(P\) and \(P\implies Q\) to obtain \(Q\), then in the second step, we use this together with \(Q\implies R\) to derive \(R\).
The inheritance rule keeps formulas in the database unless there is a contradiction:
| \(t\): | \(P, not\_know(\neg P), \not\vdash P=Now(t)\) |
| \(t+1\): | \(P\) |
| \(t\): | \(P, not\_know(P)\) |
| \(t+1\): | \(\neg P\) |
\(not\_know(P)\) is true iff \(P\) is not in the current database. Since the database is finite, this poses no computational problems. \(\not\vdash P = Now(t)\) verifies that \(P\) is not of the form \(Now(t)\) and prevents time form being inherited.
This pair of rules causes contradictions not to be present after they are detected.
Let the sentences initially present in the database be: $$Now(0), Bird(tweety), Bird(x) \land not\_know(\neg fly(x)) \implies fly(x)$$
With the above rules of inference, this is what the database looks like at consecutive steps:
At step 0: $$Now(0), Bird(tweety), Bird(x) \land not\_know(\neg fly(x)) \implies fly(x)$$
At step 1: $$Now(1), Bird(tweety), Bird(x) \land not\_know(\neg fly(x)) \implies fly(x),fly(tweety)$$
since \(\neg fly(tweety)\) is not present in the database at step 0.
The database will not change thereafter.
Now assume that the initial set of sentences is: $$Now(0), Bird(tweety), Bird(x) \land not\_know(\neg fly(x)) \implies fly(x),\neg fly(tweety)$$
This time we get:
At step 0: $$Now(0), Bird(tweety), Bird(x) \land not\_know(\neg fly(x)) \implies fly(x),\neg fly(tweety)$$
At step 1: $$Now(1), Bird(tweety), Bird(x) \land not\_know(\neg fly(x)) \implies fly(x),\neg fly(tweety)$$
This time, we cannot conclude that tweety flies since we know he doesn't, i.e., \(not\_know(\neg fly(tweety))\) fails since \(\neg fly(tweety)\) is present at step 0.