∧ hasDisease(x,y) ∧ BreastCancerTypes(y) ∧ TamoxifenTypes(z)
⇒ hasPosIntent(x,z)
so that we can then develop an argument such as: hBMJ1999∗,hasPosIntent(MsJones,TamE5YrCourse)i
to argue about whether or not we should give 5 years worth of tamoxifen.
However, as I noted in Secn. 6 of the last chapter, this allows us to argue about the intention to give a treatment, but not its effects, because the rules describing the effects of taking (for example) 5 years of tamoxifen have not been satisfied by the claim about intentions. To do that, we would need an argument of the form:
hr1,hasTreatment(MsJones, TamE5YrCourse)i
the claim of which would then satisfy the body ofBMJ1999in the example above. This would then let us use our existing knowledge to see what the effect of those actions would be, without having to re-write all the knowledge in our domain.
5.2
Committed Arguments
So far, ergonic arguments argue about an intention to perform some action. What I now want to do is to assume that an argument for an intention to do x is an argument for believing that x has been done. However, since these new arguments, which I shall refer to as committed arguments, are based upon the fact that we have an argument for the intention to perform an action in the first place, it would seem sensible to ensure that there remains some link between them. As we shall see, this is a matter of the presence of a sub-argument, and in this case is relatively easy to ensure.
I start by recalling that an ergonic argument is of the form hA, hasPosIntent(t1, t2)i (resp.hA0, hasNegIntent(t1, t2)i). My aim is to be able to derive an argument of the form hB,hasTreatment(c,d)i
(resp. hB,¬hasTreatment(c,d)i. Consider some argument a1= hA,hasPosIntent(t1, t2)i. Then we want
to make an argument of the form b1= hB,hasTreatment(t1, t2)i. The problem with this approach is that
there is no clear link between a1and b1; therefore, if we later decide that a1is not warranted, that has no
effect on b1. This could lead us into a situation where we had many arguments about a treatment and its
effects, without having any warranted arguments for the intention to perform the action itself. In order to avoid this problem, I develop a new type of defeasible rule, and add this to the support of a1in order
to develop b1.
Definition 5.2.1. Positive Commital Rules Let K be a vocabulary, I be the set of concrete individu- als in K, and RK be the set of defeasible rules. Let t1, t2∈ I be concrete individuals such that t1 ∈
P eopleN ames(Ω) and t2∈ T reatmentT ypes(Ω). Then a defeasible rule r ∈ RKis a positive commi-
tal rule iff it is of the formhasPosIntent(t1, t2)⇒hasTreatment(t1,t2).The set of all positive commital
5.2. Committed Arguments 98 Similarly, I define negative commital rules.
Definition 5.2.2. Negative Commital Rules Let K be a vocabulary, I be the set of concrete individuals in K, and RKbe the set of defeasible rules. Let t1, t2∈ I be two concrete individuals such that t1∈
P eopleN ames(Ω) and t2∈ T reatmentT ypes(t2) hold. Then a defeasible rule r ∈ RKis a negative
commital rule iff it is of the formhasNegIntent(t1, t2)⇒ ¬hasTreatment(t1, t2).The set of all negative
commital rules is denoted RCK-.
It will be useful to refer to both positive and negative commital rules.
Definition 5.2.3. A commital rule is either a positive commital rule or a negative commital rule. The set of all commital rules is denoted RCK, such that RCK= RCK+∪ RC-
K.
Example 5.2.4. The following are both commital rules:
hasPosIntent(MsJones, TamE5YrCourse)⇒hasTreatment(MsJones, TamE5YrCourse) hasNegIntent(MsJones, TamD2YrCourse)⇒ ¬hasTreatment(MsJones, TamD2YrCourse) My intention is to use commital rules to help develop new arguments by adding a commital rule to the support of an ergonic argument; this then creates an argument whose claim is that someone is or is not taking a treatment. Such arguments are termed committed arguments.
Definition 5.2.5. Committed Argument
Let K be a vocabulary and RKbe the set of defeasible rules. Let rcbe some commital rule in RCK.
Then an argument hA1, φ1i is a commited argument iff :
1. hA1, φ1i is an argument
2. There exists an ergonic argument hA2, φ2i that is a proper subargument of hA1, φ1i
3. Body(rc) = Claim(hA2, φ2i)
4. A1= A2∪ {rc}
The set of all committed arguments is denoted ACKsuch that ACK⊆ AK
Now we have a definition for the committed arguments, I define a function to produce them: Definition 5.2.6. Let K be some vocabulary, I be the set of concrete individuals, t1, t2be elements of
I and RKthe set of defeasible rules. Let e ∈ AR be some ergonic argument, r+c ∈ RCK +
be a positive commital rule and rc−∈ RCK
−
a negative commital rule, such that Body(rc+) = Claim(e) and Body(r−c)
= Claim(e). The Commit : AR· 7→ AC
· function takes an ergonic argument and returns a commited
one, so that for some ergonic argument e:
If e is of the form h{r1...rn},hasPosIntent(t1,t2)i, then Commit(e), is of the form
h{r1...rn} ∪ {r+c},hasTreatment(t1,t2)i
5.3. Hypothetical Arguments 99