assignment7 tweaks

[lambda.git] / hints / assignment_7_hint_4.mdwn

*       At the top of p. 13 (this is in between defs 2.8 and 2.9), GS&V give two examples, one for \[[∃xPx]] and the other for \[[Qx]]. In fact it will be most natural to break \[[∃xPx]] into two pieces, \[[∃x]] and \[[Px]]. But first we need to get clear on expressions like \[[Px]]. 

*       GS&V say that the effect of updating an information state `s` with the meaning of "Qx" should be to eliminate possibilities in which the entity associated with the peg associated with the variable `x` does not have the property Q. In other words, if we let `q` be a function from entities to `bool`s, `s` updated with \[[Qx]] should be `s` filtered by the function `fun (r, h) -> let obj = List.nth h (r 'x') in q obj`. When `... q obj` evaluates to `true`, that `(r, h)` pair is retained, else it is discarded.

        OK, we face two questions then. First, how do we carry this over to our present framework, where we're working with sets of `dpm`s instead of sets of discourse possibilities? And second, how do we decompose the behavior here ascribed to \[[Qx]] into some meaning for "Q" and a different meaning for "x"?

*       Answering the first question: we assume we've got some `bool dpm set` to start with. I won't call this `s` because that's what GS&V use for sets of discourse possibilities, and we don't want to confuse discourse possibilities with `dpm`s. Instead I'll call it `u`. Now what we want to do with `u` is to map each `dpm` it gives us to one that results in `(true, r, h)` only when the entity that `r` and `h` associate with variable `x` has the property Q. I'll assume we have some function q to start with that maps entities to `bool`s. 

        Then what we want is something like this:

                let eliminate_non_Qxs = (fun truth_value ->
                        fun (r, h) ->
                                let truth_value' =
                                        if truth_value
                                        then let obj = List.nth h (r 'x') in q obj
                                        else false
                                in (truth_value', r, h))
                in bind_set u (fun one_dpm -> unit_set (bind_dpm one_dpm eliminate_non_Qxs))

        The first seven lines here just perfom the operation we described: return a `bool dpm` computation that only yields `true` when its input `(r, h)` associates variable `x` with the right sort of entity. The last line performs the `bind_set` operation. This works by taking each `dpm` in the set and returning a `unit_set` of a filtered `dpm`. The definition of `bind_set` takes care of collecting together all of the `unit_set`s that result for each different set element we started with.

        We can call the `(fun one_dpm -> ...)` part \[[Qx]] and then updating `u` with \[[Qx]] will be:

                bind_set u \[[Qx]]

        or as it's written using Haskell's infix notation for bind:

                u >>= \[[Qx]]

*       Now our second question: how do we decompose the behavior here ascribed to \[[Qx]] into some meaning for "Q" and a different meaning for "x"?

        Well, we already know that \[[x]] will be a kind of computation that takes an assignment function `r` and store `h` as input. It will look up the entity that those two together associate with the variable `x`. So we can treat \[[x]] as an `entity dpm`. We don't worry here about sets of `dpm`s; we'll leave that to our predicates to interface with. We'll just make \[[x]] be a single `entity dpm`. Then what we want is:

                let getx = fun (r, h) ->
                        let obj = List.nth h (r 'x')
                        in (obj, r, h);;

*       Now what do we do with predicates? As before, we suppose we have a function q that maps entities to `bool`s. We want to turn it into a function that maps `entity dpm`s to `bool dpm`s. Eventually we'll need to operate not just on single `dpm`s but on sets of them, but first things first. We'll begin by lifting q into a function that takes `entity dpm`s as arguments and returns `bool dpm`s:

                fun entity_dpm -> bind_dpm entity_dpm (fun e -> unit_dpm (q e))

        Now we have to transform this into a function that again takes single `entity dpm`s as arguments, but now returns a `bool dpm set`. This is easily done with `unit_set`:

                fun entity_dpm -> unit_set (bind_dpm entity_dpm (fun e -> unit_dpm (q e)))

        Finally, we realize that we're going to have a set of `bool dpm`s to start with, and we need to compose \[[Qx]] with them. We don't want any of the monadic values in the set that wrap `false` to become `true`; instead, we want to apply a filter that checks whether values that formerly wrapped `true` should still continue to do so.

        This could be handled like this:

                fun entity_dpm ->
                        let eliminate_non_Qxs = fun truth_value ->
                                if truth_value = false
                                then unit_dpm false
                                else bind_dpm entity_dpm (fun e -> unit_dpm (q e))
                        in fun one_dpm -> unit_set (bind_dpm one_dpm eliminate_non_Qxs)

        Applied to an `entity_dpm`, that yields a function that we can bind to a `bool dpm set` and that will transform the doubly-wrapped `bool` into a new `bool dpm set`.

        If we let that be \[[Q]], then \[[Q]] \[[x]] would be:

                let getx = fun (r, h) ->
                        let obj = List.nth h (r 'x')
                        in (obj, r, h)
                in let entity_dpm = getx
                in let eliminate_non_Qxs = fun truth_value ->
                        if truth_value = false
                        then unit_dpm false
                        else bind_dpm entity_dpm (fun e -> unit_dpm (q e))
                in fun one_dpm -> unit_set (bind_dpm one_dpm eliminate_non_Qxs)

        or, simplifying:

                let getx = fun (r, h) ->
                        let obj = List.nth h (r 'x')
                        in (obj, r, h)
                in let eliminate_non_Qxs = fun truth_value ->
                        if truth_value
                        then bind_dpm getx (fun e -> unit_dpm (q e))
                        else unit_dpm false
                in fun one_dpm -> unit_set (bind_dpm one_dpm eliminate_non_Qxs)

        unpacking the definition of `bind_dpm`, that is:

                let getx = fun (r, h) ->
                        let obj = List.nth h (r 'x')
                        in (obj, r, h)
                in let eliminate_non_Qxs = fun truth_value ->
                        if truth_value
                        then (fun (r, h) ->
                                        let (a, r', h') = getx (r, h)
                                        in let u' = (fun e -> unit_dpm (q e)) a
                                        in u' (r', h')
                        ) else unit_dpm false
                in fun one_dpm -> unit_set (bind_dpm one_dpm eliminate_non_Qxs)

        continuing to simplify:

                let eliminate_non_Qxs = fun truth_value ->
                        if truth_value
                        then (fun (r, h) ->
                                        let obj = List.nth h (r 'x')
                                        let (a, r', h') = (obj, r, h)
                                        in let u' = (fun e -> unit_dpm (q e)) a
                                        in u' (r', h')
                        ) else unit_dpm false
                in fun one_dpm -> unit_set (bind_dpm one_dpm eliminate_non_Qxs)

                let eliminate_non_Qxs = fun truth_value ->
                        if truth_value
                        then (fun (r, h) ->
                                        let obj = List.nth h (r 'x')
                                        in let u' = unit_dpm (q obj)
                                        in u' (r, h)
                        ) else unit_dpm false
                in fun one_dpm -> unit_set (bind_dpm one_dpm eliminate_non_Qxs)

                let eliminate_non_Qxs = fun truth_value ->
                        if truth_value
                        then (fun (r, h) ->
                                        let obj = List.nth h (r 'x')
                                        in (q obj, r, h)
                        ) else unit_dpm false
                in fun one_dpm -> unit_set (bind_dpm one_dpm eliminate_non_Qxs)

        This is a function that takes a `bool dpm` as input and returns a `bool dpm set` as output.

        (Compare to the \[[Qx]] we had before:

                let eliminate_non_Qxs = (fun truth_value ->
                        fun (r, h) ->
                                let truth_value' =
                                        if truth_value
                                        then let obj = List.nth h (r 'x') in q obj
                                        else false
                                in (truth_value', r, h))
                in fun one_dpm -> unit_set (bind_dpm one_dpm eliminate_non_Qxs)

        Can you persuade yourself that these are equivalent?)   

*       Reviewing: now we've determined how to define \[[Q]] and \[[x]] such that \[[Qx]] can be the result of applying the function \[[Q]] to the `entity dpm` \[[x]]. And \[[Qx]] in turn is now a function that takes a `bool dpm` as input and returns a `bool dpm set` as output. We compose this with a `bool dpm set` we already have on hand:

                bind_set u \[[Qx]]

        or:

        <pre><code>u >>=<sub>set</sub> \[[Qx]]
        </code></pre>

*       Can you figure out how to handle \[[&exist;x]] on your own? If not, here are some [more hints](/hints/assignment_7_hint_5). But try to get as far as you can on your own.