plans concretized as dispositions

[Bill Duncan]

(Apologies for cross-posting)

My general question is: What does it mean for x to be concretized as y?

The RO definition of concretizes only talks about this in terms of dependence between the GDC and SDC:

Definition for concretizes (RO_0000059): A relationship between a specifically dependent continuant and a generically dependent continuant, in which the generically dependent continuant depends on some independent continuant in virtue of the fact that the specifically dependent continuant also depends on that same independent continuant. Multiple specifically dependent continuants can concretize the same generically dependent continuant.

Some in the OBO community want to relate plan specifications to the thing in peoples' heads, presumably a plan, that is causally relevant to carrying out the plan specification. I agree that it is useful to relate plan specifications to the agents that carry them out, but I am unsure concretizing plan specifications as dispositions provides a satisfactory solution that adds any “ontological” value.

I do not know enough about how the brain works to evaluate how (or if) GDCs concretize in the brain, but let's take a simpler model of computer (and its CPU).

The software is a type of plan specification. Now suppose I write a python program with the simple instruction print("hello world 1"). Under the current proposal, it seems the computer would:

- concretize the plan specification print("hello world 1") as a disposition d1.
- then d1 would be realized in some planned process.

I have five concerns about this.

1. The BFO-ISO release (which I am told will be open source) elucidates the concretizes at relation as follows:
Elucidation: an s-dependent continuant or process b concretizes a g-dependent continuant c at t when c is the pattern or content which b shares with actual or potential copies [207-BFO]
A disposition is definitely not a pattern, and I am skeptical that a disposition can be thought of as 'content' in any meaningful way.

2. For every instruction issued to the computer, a new disposition would be created. E.g., print("hello word 2"), print("hello word 3"), print("hello word 4"), ... would each be a separate disposition that comes to be in computer. I find this suspicious. It seems more reasonable to model this as:

- The computer has a particular disposition to print strings to standard output.
- A particular string concretizes a GDC (e.g., "hello world X") as electronic/magnetic pattern (i.e., a particular quality).
- A particular process realizes the print disposition that reads the string.

3. We can end up with a kind of mirroring between plan specifications and dispositions. That is, every plan specification will have a corresponding disposition. This could lead to a lot of extra work.

4. How do we account for failures to execute plan specifications? Dispositions are binary in their realizations. The either fire or don't. If I issue the command print("hello world 1"), and the computer prints "foo bar", I want to say the computer failed to execute the command. Perhaps we can say the plan specification failed to be concretized, but representing this may be complicated.

5. Many plan specifications/protocols involve multiple agents. An agent may understand the protocol as a whole, but only be responsible for certain parts. Does the agent have a disposition that concretizes the whole protocol, the part the agent is responsible for, or two dispositions (one for the whole protocol and one for the part)? I'm not saying that it is not possible to account for this by concretizing dispositions, I'm just worried it might be overly obtuse.

 As a counter proposal consider the following account for the aforementioned computer example:

- Computer c1 has particular disposition d1 to print strings to standard output.
- GDC g1 is an instance of the plan specification print("hello world").
- String s1 is a particular electromagnetic quality/pattern that concretizes g1.
- The memory location m1 is the particular location that bears s1.
- The particular planned process p1 realizes d1, has specified input m1, and has specified output o1.
- o1 is the set of qualities and continuants that represent a particular instance of "hello world" displaying on the screen.

This counter proposal addresses my concerns raised above.

- Regarding #1 & #2: We only have a general disposition to read/execute instructions. Not a disposition for each instruction. Depending on the level of granularity you want, you may end up mirroring of the qualities (and associated continuants) that concretize the plan specification. In some cases this is useful; e.g., you want to track a letter or number of copies printed. In others you can simply use anonymous individuals.

- Regarding #3: We can establish relations between the process that carries out the plan specification and the plan specification itself; e.g.; a particular process is prescribed by a particular plan specification, or a the process adheres to a plan specification. The determination of whether a particular process adheres to plan specification can, if necessary, be the output of some evaluation process. I'm thinking of cases in which you need to validate experimental results.

- Regarding #4: We can more easily distinguish between specification parts. Agent 1 executes a process that is prescribed by part 1 of the specification. Agent 2 executes a process that is prescribed by part 2 of the specification.

Ok ... I've said a lot (perhaps too much). What are your thoughts?

[Barry Smith]

BD writes; 3. We can end up with a kind of mirroring between plan specifications and dispositions. That is, every plan specification will have a corresponding disposition. This could lead to a lot of extra work.  

There will be many plan specifications that never correspond to a disposition.

And many of those which do correspond will be such that the disposition is never realized.

E.g. a plan specification to get rich at poker which no one ever believes in enough to become disposed to realize it.

How does your counter-proposal -- which I am sympathetic to -- affect the definition of 'concretization'?
BS

[Pierre Grenon]

Hi Bill,

I find this hard to follow. I don't think the computer analogy is simpler than the original scenario, which can be only illustrative at this level anyway. (There's loads of component dispositions and concretisations of whatnots involved in running a set of interpreted instructions...)

What precisely is the position you are trying to counter? That plan specifications are concretised in people's head at all or that their are concretised as dispositions? Is one model of your idea that plan specifications are concretised as neuron stuff in some state? And then is your argument that dispositions are not neuron stuff in some state?

I don't find that there is anything in BFO 2, other than language, that would prevent one from entertaining the notion that plan specifications are concretised as disposition. Surely, I find it a bit weird and counterintuitive because there is an element of intentionality to a plan that I'm not too sure BFO knows how to handle in terms of dispositions---it seems that the closest would be to talk about some form of cognition (has anybody said anything about this recently?). You can probably construct examples of there being a plan and no disposition or there being a disposition and no plan, etc. But this seems to be on the basis of an intuitive understanding of dispositions that I have not seen formalised in BFO yet --- Barry?

So, if the view is that concretisations of plan specifications are encoded as physical states, that bit does not seem at all incompatible with the definition of disposition (in BFO 2) and so dispositions remain candidates. If BFO can define plans better and distinguish them from dispositions, then that is another matter.

Concretisation should be used to express the categorial match between certain GDCs and corresponding SDCs. The elucidation of concretisation is weaker than that because it is in fact the job of concretisation to say that an instance of some class and not another is a concretisation of a GDC. What you could add to concretisation is that the relata are somehow categorially relevant but then this is the elusive element that makes it so that the relation is used in the first place.

Of course, if we are talking about concretization, all bets are open...

Best,

Pierre

--
You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAM2Eo-crNG_L8XKQKak2XP%3DkdTwLF4KEpGWHUQZD_Ck%3Difx0fA%40mail.gmail.com.

[Alan Ruttenberg]

I'm going to apologize in advance if the tone isn't relaxed enough.  I'm trying to get something out before I get distracted by something else :-)

Hopefully enough thoughts to be of some use, though.

On Tue, Nov 12, 2019 at 5:44 PM Bill Duncan <wddu...@gmail.com> wrote:

(Apologies for cross-posting)

FYI I only see the bfo-discuss thread and am responding to it, so no visible cross-posting here.

My general question is: What does it mean for x to be concretized as y?

A fine question. I don't think we've made it clear enough because I don't think we're clear enough about it.

The original source for the term, Ingarden, has concretizations not being copies, as we usually do, but personal adaptations.

We haven't explored what the boundaries of concretizations are, even though we tend to be loose about its use. For example, is a translation of a document a concretization? When is a version another concretization vs being a new GDC?

The RO definition of concretizes only talks about this in terms of dependence between the GDC and SDC:

Definition for concretizes (RO_0000059): A relationship between a specifically dependent continuant and a generically dependent continuant, in which the generically dependent continuant depends on some independent continuant in virtue of the fact that the specifically dependent continuant also depends on that same independent continuant. Multiple specifically dependent continuants can concretize the same generically dependent continuant.

Some in the OBO community want to relate plan specifications to the thing in peoples' heads, presumably a plan, that is causally relevant to carrying out the plan specification. I agree that it is useful to relate plan specifications to the agents that carry them out, but I am unsure concretizing plan specifications as dispositions provides a satisfactory solution that adds any “ontological” value.

 

Some in the OBO community have. The state you describe was put into practice years ago.  It brings value because otherwise there is no clear relation between some information and some action.  It makes some sense and doesn't require inventing yet another way to say it.

I do not know enough about how the brain works to evaluate how (or if) GDCs concretize in the brain, but let's take a simpler model of computer (and its CPU).

It's clear that I can memorize things, and it isn't clear at all what the inner workings of the brain are. What is clear is that there's something about me once I've memorized something that is realized in me e.g. reciting it. Whatever I have is manifested by a correlated process, if it is manifested. That sounds like a disposition. That doesn't mean there isn't more. Seems to me that we are on safer ground claiming a disposition than anything else that requires a more detailed understanding of the brain.

The software is a type of plan specification. Now suppose I write a python program with the simple instruction print("hello world 1"). Under the current proposal, it seems the computer would:

- concretize the plan specification print("hello world 1") as a disposition d1.
- then d1 would be realized in some planned process.

I have five concerns about this.

1. The BFO-ISO release (which I am told will be open source) elucidates the concretizes at relation as follows:
Elucidation: an s-dependent continuant or process b concretizes a g-dependent continuant c at t when c is the pattern or content which b shares with actual or potential copies [207-BFO]
A disposition is definitely not a pattern, and I am skeptical that a disposition can be thought of as 'content' in any meaningful way.

Well, it's an elucidation.

I'm not sure how you are so sure about disposition while accepting that qualities are patterns. Neither is manifested other than by their material (or occurrent) incarnations. So is it only material that can have patterns? And if not, please explain why qualities are and dispositions are not. And why can qualities be thought of as 'content' but dispositions not?

2. For every instruction issued to the computer, a new disposition would be created. E.g., print("hello word 2"), print("hello word 3"), print("hello word 4"), ... would each be a separate disposition that comes to be in computer. I find this suspicious. It seems more reasonable to model this as:

- The computer has a particular disposition to print strings to standard output.
- A particular string concretizes a GDC (e.g., "hello world X") as electronic/magnetic pattern (i.e., a particular quality).
- A particular process realizes the print disposition that reads the string.

3. We can end up with a kind of mirroring between plan specifications and dispositions. That is, every plan specification will have a corresponding disposition. This could lead to a lot of extra work.

There are many cases where there's a lot more to say(represent) than what we do. Part of this is handled by implication, and part that we make choices about what we need to represent because we can't represent everything in the world. There are a heck of a lot relational qualities that inhere in a body but we don't have any plans to enumerate them all. I've come, over time, to accept that there is an abundance of entities in the world, and that we necessarily have to choose what we are going to represent.

4. How do we account for failures to execute plan specifications? Dispositions are binary in their realizations. The either fire or don't. If I issue the command print("hello world 1"), and the computer prints "foo bar", I want to say the computer failed to execute the command. Perhaps we can say the plan specification failed to be concretized, but representing this may be complicated.

I think this is interesting topic, but I don't think it is specific to concretizing. Same issue with function and dysfunction.

5. Many plan specifications/protocols involve multiple agents. An agent may understand the protocol as a whole, but only be responsible for certain parts. Does the agent have a disposition that concretizes the whole protocol, the part the agent is responsible for, or two dispositions (one for the whole protocol and one for the part)? I'm not saying that it is not possible to account for this by concretizing dispositions, I'm just worried it might be overly obtuse.

We have the same issue with roles. Doesn't seem specific to dispositions, nor to concretizing.

The issue with wholes vs parts is an open one when it comes to GDCs, and while BFO doesn't have a mereology for specific dependents others have proposed such - I think Andrien Barton. That there is ongoing work reinforces that that we don't have the full story yet.

 As a counter proposal consider the following account for the aforementioned computer example:

- Computer c1 has particular disposition d1 to print strings to standard output.

Seems that there is arbitrary choice of level. Why the disposition to print strings to standard output vs the disposition to print strings to a buffer with certain streams having the disposition to render on stdout. Why not a disposition to send characters to stdout. Why not the instruction set? The microcode, the hardware built-in cascades? Why aren't the only dispositions those related to the properties of the materials on a chip that determine how electrons move? Seems to me that if you can move up from the material dispositions (why transistors work) to a higher level of abstraction (instructions of any sort) then there's nothing to prevent you from keeping on abstracting upwards. In any case, representing this (there's a lot), and its all real "could lead to a lot of extra work".

- GDC g1 is an instance of the plan specification print("hello world").

Are there one or two things here: As written there are two - g1, and 'the plan specification print("hello world").'. The latter is written as a class, as the relation in the sentence is instance-of.

Maybe better: GDC g1 is the plan specification 'print("hello world")'. Or to be more precise, maybe, GDC g1 is the plan specification, a concretization of which is: 'print("hello world")'

- String s1 is a particular electromagnetic quality/pattern that concretizes g1.

 

We all handwave here. We're not clear whether all electromagnetic stuff is quality vs disposition vs process.  Its because of this indeterminacy that I've fought to have the possibilities for concretizations be broad. That way I can just say it's concretized by something and move on.

 

- The memory location m1 is the particular location that bears s1.

 

is location a site? I'm presuming you mean some physical material that bears a quality that is the concretization. Knowing how computers work, however, kind of destroys this simple notion.

1) Dynamic Ram - depends on a process too

2) Copying - contents of memory are constantly being copied / moved, etc. In memory due to GC or allocation, on flash because writing something "nearby" necessarily rewrites the neighbors too, and usually in a different place (wear leveling)

3) Registers and internal  states of all sorts. Where do you draw the boundary. In order for that bit to be executed there's lots of movement among registers and its entirely plausible that the  device/CPU doesn't have at one time a complete copy of the ram in question while having several copies of some piece of it. Regarding representing all this see "could lead to a lot of extra work".

- The particular planned process p1 realizes d1, has specified input m1, and has specified output o1.

yup.

- o1 is the set of qualities and continuants that represent a particular instance of "hello world" displaying on the screen.

Not at all clear there's only qualities involved. Requires a deep understanding of the display technology if we want to get the story right.

This counter proposal addresses my concerns raised above.

- Regarding #1 & #2: We only have a general disposition to read/execute instructions. Not a disposition for each instruction.

Wait, didn't your story say: "Computer c1 has particular disposition d1 to print strings to standard output" There are a whole lot of instructions to read an execute if you want to accomplish that.

Depending on the level of granularity

Granularity. Now that's something we *really* need more work on.

you want, you may end up mirroring of the qualities (and associated continuants) that concretize the plan specification. In some cases this is useful; e.g., you want to track a letter or number of copies printed. In others you can simply use anonymous individuals.

 

The concretizing dispositions that you're worrying about can also be anonymous individuals, or better existentially implied.

- Regarding #3: We can establish relations between the process that carries out the plan specification and the plan specification itself; e.g.; a particular process is prescribed by a particular plan specification,

concretizes o realizes

or a the process adheres to a plan specification.

achieves planned objective

The determination of whether a particular process adheres to plan specification can, if necessary, be the output of some evaluation process. I'm thinking of cases in which you need to validate experimental results.

Ok, but how is this relevant? Here's we're talking about a process (evaluation) that involves making measurements on another process (the plan specification) and relating it to a plan specification. That concretizations can be dispositions doesn't mean they aren't concretized in other ways.

- Regarding #4: We can more easily distinguish between specification parts. Agent 1 executes a process that is prescribed by part 1 of the specification. Agent 2 executes a process that is prescribed by part 2 of the specification.

If part 1 prescribes some process, then that's sufficient to be a disposition. Plan specifications have natural parts - the actions and subgoals, perhaps dispositions mirror it, along the line of Adrien's mereology.

There's got to be *some* way in which the part relates to the process. Using disposition/concretization means we can use something we already (to some extent ) understand, as opposed to inventing something new that perhaps even fewer people understand (or remember). It's useful to only have to invent new relations when one has to.

Ok ... I've said a lot (perhaps too much).

Nah :-)

What are your thoughts?

The bottom line, from my point of view, is that concretization by dispositions makes as much sense as many other things we say, and it allows us to build on top of what we already have. It's easy to see that there's lots of potential to develop this area - plenty of work to go around. The question of whether concretizations can be dispositions or not hardly registers a blip against that. I don't see how changing it one way or another is going to help us make progress on the open issues, but I have noted where having it does address *some* issues of interest.

Best,

Alan

[Bill Duncan]

Which definition of concretization? The current one or the BFO-ISO one? As far as I can tell it is more in line with the BFO-ISO definition.

To answer your question about the plan to "get rich playing poker": If you play poker but do not get rich, are realizing a "play poker to get rich" disposition?

I would say that the plan specification is to "play poker" and the goal (or objective) specification is to "get rich".

--

You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAN_82SSiLHZfHu-PnfUzUzNtvdkYp234GXip0t3jU5MmFrT_CQ%40mail.gmail.com.

[Bill Duncan]

On Tue, Nov 12, 2019 at 7:56 PM Alan Ruttenberg <alanrut...@gmail.com> wrote:

I'm going to apologize in advance if the tone isn't relaxed enough.  I'm trying to get something out before I get distracted by something else :-)

Hopefully enough thoughts to be of some use, though.

On Tue, Nov 12, 2019 at 5:44 PM Bill Duncan <wddu...@gmail.com> wrote:

(Apologies for cross-posting)

FYI I only see the bfo-discuss thread and am responding to it, so no visible cross-posting here.

My general question is: What does it mean for x to be concretized as y?

A fine question. I don't think we've made it clear enough because I don't think we're clear enough about it.

The original source for the term, Ingarden, has concretizations not being copies, as we usually do, but personal adaptations.

We haven't explored what the boundaries of concretizations are, even though we tend to be loose about its use. For example, is a translation of a document a concretization? When is a version another concretization vs being a new GDC?

Yes. These are problems with GDCs. Notice that at least in the case of translation, we have some process that we can identify, and patterns of qualities that we can compare. In the case of dispositions, we can only compare how dispositions are realized. 

The RO definition of concretizes only talks about this in terms of dependence between the GDC and SDC:

Definition for concretizes (RO_0000059): A relationship between a specifically dependent continuant and a generically dependent continuant, in which the generically dependent continuant depends on some independent continuant in virtue of the fact that the specifically dependent continuant also depends on that same independent continuant. Multiple specifically dependent continuants can concretize the same generically dependent continuant.

Some in the OBO community want to relate plan specifications to the thing in peoples' heads, presumably a plan, that is causally relevant to carrying out the plan specification. I agree that it is useful to relate plan specifications to the agents that carry them out, but I am unsure concretizing plan specifications as dispositions provides a satisfactory solution that adds any “ontological” value.

 

Some in the OBO community have. The state you describe was put into practice years ago.  It brings value because otherwise there is no clear relation between some information and some action.  It makes some sense and doesn't require inventing yet another way to say it.

I realize this has been in practice for quite some time. My goal is to get future ontologies to implement the relation between plan specifications and agents that carry them out more carefully.

 

I do not know enough about how the brain works to evaluate how (or if) GDCs concretize in the brain, but let's take a simpler model of computer (and its CPU).

It's clear that I can memorize things, and it isn't clear at all what the inner workings of the brain are. What is clear is that there's something about me once I've memorized something that is realized in me e.g. reciting it. Whatever I have is manifested by a correlated process, if it is manifested. That sounds like a disposition. That doesn't mean there isn't more. Seems to me that we are on safer ground claiming a disposition than anything else that requires a more detailed understanding of the brain.

We need to distinguish between the process of remembering and the dispositions in the brain that allow us to activate processes to retrieve information. I do concede that there very well may cases of plans specifications that do embed themselves in the nervous system as dispositions (or something like them). I'm thinking about cases of muscle memory (e.g., learning to type) or perhaps learning a new language. However, I think we should think more carefully about asserting that every plan specification (such as a study protocol) embeds itself in the agent as a disposition. I think it makes more sense to say that the plan specification is represented in the agents brain and the agent has the disposition carry out the plan. 

The software is a type of plan specification. Now suppose I write a python program with the simple instruction print("hello world 1"). Under the current proposal, it seems the computer would:

- concretize the plan specification print("hello world 1") as a disposition d1.
- then d1 would be realized in some planned process.

I have five concerns about this.

1. The BFO-ISO release (which I am told will be open source) elucidates the concretizes at relation as follows:
Elucidation: an s-dependent continuant or process b concretizes a g-dependent continuant c at t when c is the pattern or content which b shares with actual or potential copies [207-BFO]
A disposition is definitely not a pattern, and I am skeptical that a disposition can be thought of as 'content' in any meaningful way.

Well, it's an elucidation.

Yes ... it is a primitive notion. 

I'm not sure how you are so sure about disposition while accepting that qualities are patterns. Neither is manifested other than by their material (or occurrent) incarnations. So is it only material that can have patterns? And if not, please explain why qualities are and dispositions are not. And why can qualities be thought of as 'content' but dispositions not?

In principle, qualities are observable, whereas disposition, in principle, are not observable. Only the processes that realize dispositions can be observed. I have a hard time making sense of the notion that something which cannot be observed can have a pattern or content. How can you copy such a thing or compare it (directly) to another? With qualities, this is more understandable. Also, I do think it is sensible for GDCs to be concretized in processes.

 

2. For every instruction issued to the computer, a new disposition would be created. E.g., print("hello word 2"), print("hello word 3"), print("hello word 4"), ... would each be a separate disposition that comes to be in computer. I find this suspicious. It seems more reasonable to model this as:

- The computer has a particular disposition to print strings to standard output.
- A particular string concretizes a GDC (e.g., "hello world X") as electronic/magnetic pattern (i.e., a particular quality).
- A particular process realizes the print disposition that reads the string.

3. We can end up with a kind of mirroring between plan specifications and dispositions. That is, every plan specification will have a corresponding disposition. This could lead to a lot of extra work.

There are many cases where there's a lot more to say(represent) than what we do. Part of this is handled by implication, and part that we make choices about what we need to represent because we can't represent everything in the world. There are a heck of a lot relational qualities that inhere in a body but we don't have any plans to enumerate them all. I've come, over time, to accept that there is an abundance of entities in the world, and that we necessarily have to choose what we are going to represent.

I agree. I'm trying to keep the story of what to enumerate short without things that may not make sense. Again, I will concede that there are cases where would want to assert that a plan specification is concretized as a disposition. Suppose I build a machine devoid of software components that plays Jingle Bells every time I press a button. In this case, the hardwired components may create in the machine a disposition to play Jingle Bells. From what I've seen, many of use cases for concretizing plans has to do with humans carrying out protocols.

  

4. How do we account for failures to execute plan specifications? Dispositions are binary in their realizations. The either fire or don't. If I issue the command print("hello world 1"), and the computer prints "foo bar", I want to say the computer failed to execute the command. Perhaps we can say the plan specification failed to be concretized, but representing this may be complicated.

I think this is interesting topic, but I don't think it is specific to concretizing. Same issue with function and dysfunction.

Not directly, but it does have to do with concretizing plan specifications as dispositions. If a plan specification is concretized as disposition, then what kind of story can we tell about cases in which the dispositions fails to follow the plan specification? Would we then assert that the plan specification was not concretized?

5. Many plan specifications/protocols involve multiple agents. An agent may understand the protocol as a whole, but only be responsible for certain parts. Does the agent have a disposition that concretizes the whole protocol, the part the agent is responsible for, or two dispositions (one for the whole protocol and one for the part)? I'm not saying that it is not possible to account for this by concretizing dispositions, I'm just worried it might be overly obtuse.

We have the same issue with roles. Doesn't seem specific to dispositions, nor to concretizing.

The issue with wholes vs parts is an open one when it comes to GDCs, and while BFO doesn't have a mereology for specific dependents others have proposed such - I think Andrien Barton. That there is ongoing work reinforces that that we don't have the full story yet.

 As a counter proposal consider the following account for the aforementioned computer example:

- Computer c1 has particular disposition d1 to print strings to standard output.

Seems that there is arbitrary choice of level. Why the disposition to print strings to standard output vs the disposition to print strings to a buffer with certain streams having the disposition to render on stdout.

Depends on the level granularity you wanting/needing to represent. I am using a high level example. If you were wanting to represent the whole IO architecture, then (of course) many more details would be needed, and there would be different architectures to represent too.

 

Why not a disposition to send characters to stdout. Why not the instruction set? The microcode, the hardware built-in cascades? Why aren't the only dispositions those related to the properties of the materials on a chip that determine how electrons move? Seems to me that if you can move up from the material dispositions (why transistors work) to a higher level of abstraction (instructions of any sort) then there's nothing to prevent you from keeping on abstracting upwards. In any case, representing this (there's a lot), and its all real "could lead to a lot of extra work".

- GDC g1 is an instance of the plan specification print("hello world").

Are there one or two things here: As written there are two - g1, and 'the plan specification print("hello world").'. The latter is written as a class, as the relation in the sentence is instance-of.

Maybe better: GDC g1 is the plan specification 'print("hello world")'. Or to be more precise, maybe, GDC g1 is the plan specification, a concretization of which is: 'print("hello world")'

- String s1 is a particular electromagnetic quality/pattern that concretizes g1.

 

We all handwave here. We're not clear whether all electromagnetic stuff is quality vs disposition vs process.  Its because of this indeterminacy that I've fought to have the possibilities for concretizations be broad. That way I can just say it's concretized by something and move on.

 

- The memory location m1 is the particular location that bears s1.

 

is location a site? I'm presuming you mean some physical material that bears a quality that is the concretization. Knowing how computers work, however, kind of destroys this simple notion.

Yes. I told a simplistic story about memory locations. In reality there is a lot complexity to how computer memory works. As noted above, BFO-ISO allows for GDCs to be concretized in processes.

 

1) Dynamic Ram - depends on a process too

2) Copying - contents of memory are constantly being copied / moved, etc. In memory due to GC or allocation, on flash because writing something "nearby" necessarily rewrites the neighbors too, and usually in a different place (wear leveling)

3) Registers and internal  states of all sorts. Where do you draw the boundary. In order for that bit to be executed there's lots of movement among registers and its entirely plausible that the  device/CPU doesn't have at one time a complete copy of the ram in question while having several copies of some piece of it. Regarding representing all this see "could lead to a lot of extra work".

- The particular planned process p1 realizes d1, has specified input m1, and has specified output o1.

yup.

- o1 is the set of qualities and continuants that represent a particular instance of "hello world" displaying on the screen.

Not at all clear there's only qualities involved. Requires a deep understanding of the display technology if we want to get the story right.

You are right. There are many processes involved and dispositions for the machine to carry out certain operations. Again, I was trying to keep the example somewhat simple. The main point, however, is just that every time a CPU adds two integers, it doesn't acquire a new disposition to add those integers. Rather the CPU has disposition to add the contents of two registers. The contents of those registers are constantly changing.

This counter proposal addresses my concerns raised above.

- Regarding #1 & #2: We only have a general disposition to read/execute instructions. Not a disposition for each instruction.

Wait, didn't your story say: "Computer c1 has particular disposition d1 to print strings to standard output" There are a whole lot of instructions to read an execute if you want to accomplish that.

Depending on the level of granularity

Granularity. Now that's something we *really* need more work on.

you want, you may end up mirroring of the qualities (and associated continuants) that concretize the plan specification. In some cases this is useful; e.g., you want to track a letter or number of copies printed. In others you can simply use anonymous individuals.

 

The concretizing dispositions that you're worrying about can also be anonymous individuals, or better existentially implied.

- Regarding #3: We can establish relations between the process that carries out the plan specification and the plan specification itself; e.g.; a particular process is prescribed by a particular plan specification,

concretizes o realizes

or a the process adheres to a plan specification.

achieves planned objective

You can achieve an objective in may different ways. Perhaps you just get lucky. There are important cases in which you want to be able able to assert that the plan specification was in fact followed and the objective was achieved.

 

The determination of whether a particular process adheres to plan specification can, if necessary, be the output of some evaluation process. I'm thinking of cases in which you need to validate experimental results.

Ok, but how is this relevant? Here's we're talking about a process (evaluation) that involves making measurements on another process (the plan specification) and relating it to a plan specification. That concretizations can be dispositions doesn't mean they aren't concretized in other ways.

- Regarding #4: We can more easily distinguish between specification parts. Agent 1 executes a process that is prescribed by part 1 of the specification. Agent 2 executes a process that is prescribed by part 2 of the specification.

If part 1 prescribes some process, then that's sufficient to be a disposition. Plan specifications have natural parts - the actions and subgoals, perhaps dispositions mirror it, along the line of Adrien's mereology.

We might be able to assert that the processes follow the plan specification, but given the non-observable nature of dispositions I think it is hard to assert (except in cases noted above) that they mirror plan specifications.

 

There's got to be *some* way in which the part relates to the process. Using disposition/concretization means we can use something we already (to some extent ) understand, as opposed to inventing something new that perhaps even fewer people understand (or remember). It's useful to only have to invent new relations when one has to.

Ok ... I've said a lot (perhaps too much).

Nah :-)

What are your thoughts?

The bottom line, from my point of view, is that concretization by dispositions makes as much sense as many other things we say, and it allows us to build on top of what we already have. It's easy to see that there's lots of potential to develop this area - plenty of work to go around. The question of whether concretizations can be dispositions or not hardly registers a blip against that. I don't see how changing it one way or another is going to help us make progress on the open issues, but I have noted where having it does address *some* issues of interest.

Thanks for your input. I appreciate it! Hopefully, I've given an adequate response :)

Best,

Alan

 

--
You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAM2Eo-crNG_L8XKQKak2XP%3DkdTwLF4KEpGWHUQZD_Ck%3Difx0fA%40mail.gmail.com.

--
You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAFKQJ8mFC6m-LAinE6o0f%2BYOSbv_O1txZwq5QpFVOqhT%3DuBXsA%40mail.gmail.com.

[Bill Duncan]

Hi Pierre,

Thanks for your response!

To be more specific, my concern is about concretizing plans as dispositions. I think in some cases, there is tendency to link agents to plan specifications using this approach, and it is not necessarily warranted. Let me know if my response to Alan addressed your concerns, and made my line of thinking clearer.

Cheers,

Bill

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAAkf34zZTPv6OegP2_pZV117xWgY7Aw0-jfnyzRu51cabjocGA%40mail.gmail.com.

[Thomas Beale]

This may be a dumb Q, but I had always thought that a plan (in the sense of things like computer programs) being concretised in an SDC meant doing so in the same sense of concretising Shakespeares sonnets (some content) into a particular book copy I go and buy at Waterstones. I.e. a physical media concept. You are describing plan 'execution'. Is this meant to come under the rubric of concretisation in BFO as well? I would have thought plan 'realisation' was execution.

 It seems to me that physical plan representation (concretisation) and execution (realisation of intention) are quite different things...

- thomas

[Pierre Grenon]

Ok, thank you. No I didn't find the answer to Alan useful but that's because I got lost. OTOH, I wouldn't go as far as saying that I have concerns.

So we assume there are plans and dispositions and we are asking whether plans, like functions, say, are dispositions. We kinda already know that they are in the same ballpark, at least cousins if not siblings in the DC family.

So what do dispositions exhibit that plans don't and reciprocally? I didn't fully get that from your answers to Alan. 

When you say relationship between plans and agents, what agent do you have in mind and at which stage of the 'plan specification' lifecycle? 

Say we look at 3 possible phases: conception, execution, contemplation (as in I contemplate a plan spec in order to communicate it or to evaluate conformance of the output of the execution of a plan that allegedly concretises the plan specification in question).

Can I conceive of a plan without having the disposition that concretises the plan specification, myself?

Can I carry out a plan without it being a disposition in accord with the plan specification?

Can I contemplate a plan without having a disposition that concretises the plan?

I'd suggest tackling these issues and going back to your initial question.

With many thanks and kind regards,
Pierre

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAM2Eo-cAmzHdi3U2qavTxHBwuLhmmLeJ31h-9gXSJV7oX8pKsA%40mail.gmail.com.

[Pierre Grenon]

It seems to me Bill is adressing the suggestion that one physical concretisation of plans is in people and then that this concretisation is a disposition. 

What in BFO enforces formally that plans are like scores? And what prevents us from saying that Shakespear's sonnets are concretised as dispositions in people? (Maybe IAO has something to say.)

I don't know, to be clear. My intuition is that along the line of your drawing attention to execution (or 'realisation', whatever), it would trivialise the notion of plan to claim that there is a disposition involved in execution and that this is the concretisation of the plan specification. 

To my mind, the challenge is to account for the cognition of agents as standardly one expects that plan following is intentional and deliberate. 

If someone were to carry out processes in an apparently orderly fashion yet for all they knew in a completely random manner, we should be reluctant to say they had a disposition to do so that concretised a plan specification in a non trivial sense. If this were our view, I don't see what would prevent us from saying that whatever people do is a concretisation of a plan specification that we can specify after the fact... and it seems odd to me to say this in the general case. 

It doesn't mean that dispositions do not exist in people who act, whether deliberately or not, only that under the view there has to be something in addition that accounts for the planning element when it is there. 

Best,

Pierre

--
You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/fd52e9f4-71cf-4a5d-a495-cae78235a069%40googlegroups.com.

[Thomas Beale]

These distinctions are not only academic, which is why I raised the question.

Consider a 'plan' that is actually a plan - say a work plan for execution of a clinical protocol on a certain kind of patient (e.g. intubation, diagnosis of angina etc). 

The plan, such as it is, exists originally (say) in the heads of some professors of medicine who want to standardise intubation for the NHS; an agreed version becomes a semi-structured representation on the Nice.org.uk website, where it is a GDC - each physically recorded, printed or displayed copy is its own SDC (this however, seems uninteresting).

Now imagine it is 'translated' to a formal representation in some plan language, so that it can be executed in a plan execution engine (a thing that will push notifications to the screens of various workers, to get them to do the tasks in the plan) and also ask them questions to find out the current state of reality.

Now, prior to that execution, there will be plan verification tools that process a / the ?concretisation of the plan that would be executed, to ensure it is safe, doesn't contain deadlock etc.

Note that plans in general contain decision points (referring to environmental variables, etc), so the representation of a plan is not in general even the same logical structure as its execution - a plan execution is one possible traversal through the decision graph of a plan. 

The now-verified plan representation might be compiled into a form directly executable by the plan engine. 

When it does execute, it can leave a detailed action log (some people call this 'digital exhaust') of the kind that can be used by process mining activities to analyse the plan as it actually did execute, i.e. what traversal was actually taken on this particular run.

As I am sure everyone here knows very well, all of these intermediate representations, executions and forensic examinations are routine IT activities these days. So I would have thought that 'concretisations' more usefully refer to representations of things, while executions are something else. If BFO/RO 'concretisation' is to encompass the latter as well, it's not yet clear to me how.

- thomas

To unsubscribe from this group and stop receiving emails from it, send an email to bfo-d...@googlegroups.com.

[Bill Duncan]

As far as I know, BFO does not formally prohibit a plan specification of being concretized as dispositions, although, BFO-ISO might have some restrictions.

I was trying to stay away from human cases b/c we have all sorts of complicated scenarios: e.g., if I have an intention to carry out a plan specification and then change my mind, was the plan specification concretized as a disposition in my brain?

Many of my concerns come about as a result of thinking about malfunctions. How can you evaluate the content of a disposition? You can evaluate the process that realized a disposition, but what about the disposition itself?

Bill

To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/621a72de-1d72-4dee-9d75-b1ca05800132%40googlegroups.com.

[Barry Smith]

Responses to Bill in-line 

I have five concerns about this.

1. The BFO-ISO release (which I am told will be open source) elucidates the concretizes at relation as follows:
Elucidation: an s-dependent continuant or process b concretizes a g-dependent continuant c at t when c is the pattern or content which b shares with actual or potential copies [207-BFO]
A disposition is definitely not a pattern, and I am skeptical that a disposition can be thought of as 'content' in any meaningful way.

I do not see how you infer from this that a disposition is a pattern

 

2. For every instruction issued to the computer, a new disposition would be created. E.g., print("hello word 2"), print("hello word 3"), print("hello word 4"), .

this seems exactly right to me; when you follow a recipe in your kitchen as you read each line you formal a corresponding disposition which you then execute in corresponding cooking actions

 

.. would each be a separate disposition that comes to be in computer. I find this suspicious. It seems more reasonable to model this as:

- The computer has a particular disposition to print strings to standard output.
- A particular string concretizes a GDC (e.g., "hello world X") as electronic/magnetic pattern (i.e., a particular quality).
- A particular process realizes the print disposition that reads the string.

when the process realizes the disposition to print "hello world X" then it thereby all realizes a disposition to print strings to standard output (the latter is not something extra, it is there just as, when you sing Jingle Bells you also realize a disposition to sing.

 

3. We can end up with a kind of mirroring between plan specifications and dispositions. That is, every plan specification will have a corresponding disposition. This could lead to a lot of extra work.

BFO is here to help us represent reality, not to help us find shortcuts.

4. How do we account for failures to execute plan specifications? Dispositions are binary in their realizations. The either fire or don't.

If the printer runs out of power after it printed only "hello wo"?

 

 

If I issue the command print("hello world 1"), and the computer prints "foo bar", I want to say the computer failed to execute the command. Perhaps we can say the plan specification failed to be concretized, but representing this may be complicated.

Life is hard.

 

5. Many plan specifications/protocols involve multiple agents. An agent may understand the protocol as a whole, but only be responsible for certain parts. Does the agent have a disposition that concretizes the whole protocol, the part the agent is responsible for, or two dispositions (one for the whole protocol and one for the part)? I'm not saying that it is not possible to account for this by concretizing dispositions, I'm just worried it might be overly obtuse.

It would be obtuse (complicated) but  only to the appropriate degree. 

d has specified output o1.
- o1 is the set of qualities and continuants that represent a particular instance of "hello world" displaying on the screen.

This counter proposal addresses my concerns raised above.

- Regarding #1 & #2: We only have a general disposition to read/execute instructions. Not a disposition for each instruction.

We acquire said disposition every time we read an instruction; same applies to the computer

 

Depending on the level of granularity you want, you may end up mirroring of the qualities (and associated continuants) that concretize the plan specification. In some cases this is useful; e.g., you want to track a letter or number of copies printed. In others you can simply use anonymous individuals.

There are two issues here: how granular do we want to description/representation to be (for a given purpose)

And: what does on on the different levels of granularity within whatever it is that is (e.g.) realizing a plan.

We need the resources to deal with the latter at all levels of granularity.

.

[Bill Duncan]

Responses to Barry in-line.

On Thu, Dec 26, 2019 at 7:16 AM Barry Smith <ifo...@gmail.com> wrote:

Responses to Bill in-line 

I have five concerns about this.

1. The BFO-ISO release (which I am told will be open source) elucidates the concretizes at relation as follows:
Elucidation: an s-dependent continuant or process b concretizes a g-dependent continuant c at t when c is the pattern or content which b shares with actual or potential copies [207-BFO]
A disposition is definitely not a pattern, and I am skeptical that a disposition can be thought of as 'content' in any meaningful way.

I do not see how you infer from this that a disposition is a pattern

What does it mean for an ICE to be concretized as a disposition then? I can make sense of an ICE concretized as quality or temporal pattern, but not a disposition.

 

2. For every instruction issued to the computer, a new disposition would be created. E.g., print("hello word 2"), print("hello word 3"), print("hello word 4"), .

this seems exactly right to me; when you follow a recipe in your kitchen as you read each line you formal a corresponding disposition which you then execute in corresponding cooking actions

This seems wrong to me. A more precise account would be that the computer has a process that prints strings to the screen (or some other device). This printing process then takes strings as inputs. A new disposition is not created for every string you print.

 

 

.. would each be a separate disposition that comes to be in computer. I find this suspicious. It seems more reasonable to model this as:

- The computer has a particular disposition to print strings to standard output.
- A particular string concretizes a GDC (e.g., "hello world X") as electronic/magnetic pattern (i.e., a particular quality).
- A particular process realizes the print disposition that reads the string.

when the process realizes the disposition to print "hello world X" then it thereby all realizes a disposition to print strings to standard output (the latter is not something extra, it is there just as, when you sing Jingle Bells you also realize a disposition to sing.

Not sure I follow you here. You seem to be agreeing with my comment above. Are you saying that when you sing "Jingle Bells", you realize your "Sing Jingle Bells" disposition? Or are you saying that you realize your disposition to sing, and that during the "singing process" you input (from memory) the word to "Jingle Bells"? I think the latter makes more sense.

 

3. We can end up with a kind of mirroring between plan specifications and dispositions. That is, every plan specification will have a corresponding disposition. This could lead to a lot of extra work.

BFO is here to help us represent reality, not to help us find shortcuts.

Agreed. I am trying the represent reality. The extra entities that are created gives me pause to assert them in the absence of scientific evidence.  

4. How do we account for failures to execute plan specifications? Dispositions are binary in their realizations. The either fire or don't.

If the printer runs out of power after it printed only "hello wo"?

So in this case we could say that the printer function was realized, but the final output did not meet our expectation. In other words we can evaluate the inputs and outputs against specifications in order to determine where or when the malfunction occurred. Dispositions do not offer this. They are empirically opaque.

 

 

 

If I issue the command print("hello world 1"), and the computer prints "foo bar", I want to say the computer failed to execute the command. Perhaps we can say the plan specification failed to be concretized, but representing this may be complicated.

Life is hard. 

 

5. Many plan specifications/protocols involve multiple agents. An agent may understand the protocol as a whole, but only be responsible for certain parts. Does the agent have a disposition that concretizes the whole protocol, the part the agent is responsible for, or two dispositions (one for the whole protocol and one for the part)? I'm not saying that it is not possible to account for this by concretizing dispositions, I'm just worried it might be overly obtuse.

It would be obtuse (complicated) but  only to the appropriate degree. 

d has specified output o1.
- o1 is the set of qualities and continuants that represent a particular instance of "hello world" displaying on the screen.

This counter proposal addresses my concerns raised above.

- Regarding #1 & #2: We only have a general disposition to read/execute instructions. Not a disposition for each instruction.

We acquire said disposition every time we read an instruction; same applies to the computer

No. The more accurate model is that the computer system has a function to execute instructions that are loaded into its registers (assuming we are talking about modern CPUs). Not that every time an every time an instruction is loaded into its registers the CPU acquires a new disposition.

 

Depending on the level of granularity you want, you may end up mirroring of the qualities (and associated continuants) that concretize the plan specification. In some cases this is useful; e.g., you want to track a letter or number of copies printed. In others you can simply use anonymous individuals.

There are two issues here: how granular do we want to description/representation to be (for a given purpose)

And: what does on on the different levels of granularity within whatever it is that is (e.g.) realizing a plan.

We need the resources to deal with the latter at all levels of granularity.

.

--

You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAN_82SRarR-Z%2B5d5p38gtCVZoyUKrfMP5bGS51AR%2BU72W0eFyQ%40mail.gmail.com.

[Victor Maojo]

Queridos colegas, estos son de aquel grupo cercano a Barry Smith, los que habíamos criticado en el artículo de 2010. Como se ve, Estabamos confundidos. La filosofía ontológica va mejorando en su precisión. 

Un abrazo, Víctor

Enviado desde mi iPhone

--

You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAM2Eo-crNG_L8XKQKak2XP%3DkdTwLF4KEpGWHUQZD_Ck%3Difx0fA%40mail.gmail.com.

[Barry Smith]

Victor writes (roughly translated):
Dear colleagues, these are from the group close to Barry Smith, the ones we had criticized in the 2010 article. As you can see, we were confused. The ontological philosophy is improving in its accuracy.

This is good news.

BS

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/E35A775A-89EC-4332-983C-B6FD618CECD3%40infomed.dia.fi.upm.es.

[Bill Duncan]

Hopefully, this example makes clear my skepticism about plan specifications being concretized as dispositions in the case of computer systems.

 

Consider an AND logic gate:

https://en.wikipedia.org/wiki/AND_gate

 

The behavior of this gate can be summarized as:

 

Given inputs A, B, and output C:

C = 1, if A = 1 and B = 1

Otherwise, C = 0

 

Let’s say this device is implemented as a physical device with A, B, and C being electronic registers that are either on (+) or off (-). Basically, “+” represents a “1”, “-“ represents a “0”, and the value of C (i.e., whether C is on or off) is determined by the +/- values of A and B.

 

If I have a simple language that consists of characters “1” and “0”, and an instruction is 2 characters long, I can write programs (i.e., sequences of instructions) that result in some value being loaded into C.

 

I agree that the physical device has dispositions (although, whether it is one complex disposition or a set of dispositions seems a bit unclear to me). However, the possession of said disposition/s by the physical device is not the same as concretizing ICEs as dispositions. Suppose, I load the instruction “10” into the device. This gets translated into A being on, B being off, and resulting in C being off. Symbolically, we could represent this behavior something like this:

A(+), B(-) -> C(-)

 

If we take the instruction to be an ICE, we have “10” concretized as A(+), B(-). The electromagnetic properties of A and B then trigger the realization of the behavior of the electronic device, which results in C(-).

 

In this scenario, it makes sense that “10” is concretized as A(+), B(-). It is a transformation of one pattern into another. The question at hand, however, is whether “10” is concretized as the disposition whose realization is described by:

A(+), B(-) -> C(-)

 

I think the answer to this question is “no”. The device already has the disposition to process (so to speak) the instruction “10”. The loading of registers A and B triggers the realization of this disposition.

 

I realize my example has been from the vantage point of a simple computational system. The human brain may function much differently.

Bill

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAN_82SSn%2BipXnBvgvxz-sYq9MrLPbQ-wfCQDs7u8uats_--%2BBg%40mail.gmail.com.

[Bill Duncan]

This thread seems to grown stale. So, I assume everyone agrees with me :)

To summarize, what has been discussed so far.

1. Many ontologies are using the concretizes to hold between a plan specification and a realizable entity. Here is how concretizes is defined in RO:

label: concretizes (http://purl.obolibrary.org/obo/RO_0000059)
Definition: A relationship between a specifically dependent continuant and a generically dependent continuant, in which the generically dependent continuant depends on some independent continuant in virtue of the fact that the specifically dependent continuant also depends on that same independent continuant. Multiple specifically dependent continuants can concretize the same generically dependent continuant.

An example of using concretizes, is found in OBI's planned process class:

label: planned process (http://purl.obolibrary.org/obo/OBI_0000011)
Definition: A processual entity that realizes a plan which is the concretization of a plan specification.

The formal definition of planned process includes the axiom:

realizes some (concretizes some 'plan specification')

The RO definition of 'concretizes' doesn't really tell us very much about the nature of what it means for a realizable to concretize a plan specification, and I have argued that this lack of informativeness leads to problems.

2. The new BFO-ISO release defines the concretizes relation as:

label: concretizes

Elucidation: an s-dependent continuant or process b concretizes a g-dependent continuant c at t when c is the pattern or content which b shares with actual or potential copies [207-BFO]

This definition makes it clear that the GDC is a sharable pattern (or content). Since realizable entities are not observable (only their manifestations are observable), it does not makes sense that a realizable entity can be a shareable pattern.

3. I will not rehash my arguments thus far. I propose it is ontologically more precise to represent concretizations of plan specifications as qualities that are specified inputs into the relevant processes (e.g., computational processes, cognitive processes, planned processes).

Axiomatically the change is minor. For instance, instead of defining OBI's planned process using the axiom:

realizes some (concretizes some 'plan specification')

you would use:

has_specfied_input some (concretizes some 'plan specification')

Please share your comments/thoughts.
Bill

[James A. Overton]

Hi Bill,

Since you asked: I don't agree with your conclusion, I haven't found your examples persuasive, and I'm confused about two points in your argument. 

1. You said:

Since realizable entities are not observable (only their manifestations are observable), it does not makes sense that a realizable entity can be a shareable pattern.

But nobody is proposing that realizable entities are GDCs. The OBI model is that a 'plan specification' (GDC) is concretized by a 'plan' (realizable entity) which is realized in a 'planned process'. The 'plan' is in the head of an investigator, and is not shared. The 'plan specification' is shared.

2. I believe I'm responsible for the definition of RO 'concretizes', which I tried to adapt from the BFO specification (pre-ISO) without logical formalism. The BFO-ISO elucidation that you quote is more informative. It also adds 'process' to the domain of 'concretizes'. So I was expecting you to argue that a 'planned process' should directly concretize a 'plan specification'. But then you propose an intermediate quality that is the specified input for the 'planned process'.

I also disagree that this is a minor change. Planned processes are the centre of OBI, that modelling has been in place for well over a decade. When considering such a major change, I want to be convinced that there's a major problem that needs to be fixed, and not just that the proposed change is "ontologically more precise".

James

On Jan 24, 2020, at 11:26, Bill Duncan <wddu...@gmail.com> wrote:



To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAM2Eo-doFmG1J9%3Dh1-JRduJPjdwCv1vwTM9WEpBrc4kSL1OcdA%40mail.gmail.com.

[Barry Smith]

The BFO 2020 elucidation of concretizes is:

 an s-dependent continuant or process b concretizes a g-dependent continuant c at t when c is the pattern or content which b shares with actual or potential copies [207-BFO] 

 Thus if there is a computer program (and if a computer program is a plan) then the computer program would be concretized either by writing it down on paper or by entering it into a computer (and creating a pattern in its hard drive) or (I guess) by some human brain understanding it and concretizing it via some complex neurological process.

If the program is entered into the computer then the computer has a disposition to execute it, and this disposition is triggered e.g. by pressing enter. Do you agree with that? 

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAM2Eo-doFmG1J9%3Dh1-JRduJPjdwCv1vwTM9WEpBrc4kSL1OcdA%40mail.gmail.com.

[Bill Duncan]

James:

I never said that realizables are GDCs, only that concretizing GDCs as realizables is problematic. The previous email threads detail why I think the notion of realizables concretizing plan specifications is ontologically imprecise.

By "minor change", I meant that it is a minor change syntactically. I know the OBI modeling has been in place for a long time. I've been advocating this change for a long time. Doing something wrong for a long time doesn't make it right. From a pragmatic perspective, you need to get a sense of how many users actually instantiate 'concretizations of plan specifications' in order to assess the impact.

Barry:

In general, I think I agree with you. It depends on whether you are using the expression 'computer program is plan' in the dispositional sense of 'plan' or the informational sense of 'plan' (i.e., plan specification).

 

I hold that a particular computer program (i.e., code embedded in some medium, such as magnetic patterns on disk) is a concretization of a plan specification. The computer system has many processes that behave in well defined ways in the presence of certain patterns (i.e., read and execution processes). Please see my previous email using an AND logic gate as an example of this.

Speaking loosely, you could say that a computer system has the disposition to execute computer programs. But, since only computer systems that are actively running, can execute computer programs, this can be misleading.

Thanks for the comments!

Bill

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAN_82SQVxVq%3DZ0thad-C4XUt6qw80VC4bPGKVpagfMmWzRSi2w%40mail.gmail.com.

[Barry Smith]

On Fri, Jan 24, 2020 at 1:24 PM Bill Duncan <wddu...@gmail.com> wrote:

James:

I never said that realizables are GDCs, only that concretizing GDCs as realizables is problematic. The previous email threads detail why I think the notion of realizables concretizing plan specifications is ontologically imprecise.

By "minor change", I meant that it is a minor change syntactically. I know the OBI modeling has been in place for a long time. I've been advocating this change for a long time. Doing something wrong for a long time doesn't make it right. From a pragmatic perspective, you need to get a sense of how many users actually instantiate 'concretizations of plan specifications' in order to assess the impact.

I think we should drop the idea of realizables as concretizations 

Barry:

In general, I think I agree with you. It depends on whether you are using the expression 'computer program is plan' in the dispositional sense of 'plan' or the informational sense of 'plan' (i.e., plan specification).

 

I hold that a particular computer program (i.e., code embedded in some medium, such as magnetic patterns on disk) is a concretization of a plan specification. The computer system has many processes that behave in well defined ways in the presence of certain patterns (i.e., read and execution processes). Please see my previous email using an AND logic gate as an example of this.

Yes 

Speaking loosely, you could say that a computer system has the disposition to execute computer programs. But, since only computer systems that are actively running, can execute computer programs, this can be misleading.

Anything can be misleading. But a computer in which program p is installed has a disposition to execute p even when it is turned off (which is triggered by turning it on and pressing enter). 

Methinks
BS 

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAM2Eo-chp0AVxMZvMifYWEUy8EdB5gbeCX820HJdzVBnhEhMKQ%40mail.gmail.com.

[Bill Duncan]

On Fri, Jan 24, 2020 at 1:30 PM Barry Smith <ifo...@gmail.com> wrote:

On Fri, Jan 24, 2020 at 1:24 PM Bill Duncan <wddu...@gmail.com> wrote:

James:

I never said that realizables are GDCs, only that concretizing GDCs as realizables is problematic. The previous email threads detail why I think the notion of realizables concretizing plan specifications is ontologically imprecise.

By "minor change", I meant that it is a minor change syntactically. I know the OBI modeling has been in place for a long time. I've been advocating this change for a long time. Doing something wrong for a long time doesn't make it right. From a pragmatic perspective, you need to get a sense of how many users actually instantiate 'concretizations of plan specifications' in order to assess the impact.

I think we should drop the idea of realizables as concretizations 

Barry:

In general, I think I agree with you. It depends on whether you are using the expression 'computer program is plan' in the dispositional sense of 'plan' or the informational sense of 'plan' (i.e., plan specification).

 

I hold that a particular computer program (i.e., code embedded in some medium, such as magnetic patterns on disk) is a concretization of a plan specification. The computer system has many processes that behave in well defined ways in the presence of certain patterns (i.e., read and execution processes). Please see my previous email using an AND logic gate as an example of this.

Yes 

Speaking loosely, you could say that a computer system has the disposition to execute computer programs. But, since only computer systems that are actively running, can execute computer programs, this can be misleading.

Anything can be misleading. But a computer in which program p is installed has a disposition to execute p even when it is turned off (which is triggered by turning it on and pressing enter). 

Methinks
BS 

Then yes .. we "agree" :)

 

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAN_82SRq8xFQ-Zgboy8HFy8vSxJh%2Bxk5YM8m5RTeuCYe5gaaKw%40mail.gmail.com.

[Alan Ruttenberg]

The idea that planned processes are realizations of concretizations that are dispositions is a minimal start, not an end of the story. What we wanted to do with OBI at the time (and this drove IAO) was to have a relation between a plan specification, as written, and the process that it specifies.  The choice was whether to invent a new relation or build on what we had, subject to it making sense, with a strong preference towards re-using relations where possible. The axiom deliberately says as little as possible - it's the bare minimum path between the specification and the process. It makes perfect sense in that the manifestation of realizables is a 'correlated' process. In fact, concretization as a disposition is an important property that separates plan specification from other information artifacts. If you read a novel, the actions you take later as a consequence of that reading are all over the place, and can much or little to do with the content. On the other hand, when we have a plan specification it is related to a rather specific kind of process.

The second aspect that pushes towards dispositions is that the plan specification is in some sense 'future pointing', which is the distinguishing feature of realizables.

To ignore that connection would have been, in my view, irresponsible.

But what didn't we say? Many things. I'm sympathetic to Pierre's comment about trivializing plans. But we didn't try to develop plans in detail. The best we could do, IMO, was stay out of the way for when further work was done, and I think we've been successful in that.

We didn't say exactly what the disposition was towards. We didn't say that the disposition is *only* realized in a planned process (well, we might have unintentionally due to OWL not being modal). We didn't say to what IC the concretization inheres. Because a plan consists in part of subgoals, which specify and end, not a manner, the disposition can't specify the details of particular realizations - the manner being unpredictable in many cases. We didn't work out a mereonomy of dispositions/realizations or talk about the precise structure of the disposition. We didn't say anything about the trigger. We didn't say anything specific about cognition. All this will be interesting and challenging to develop, and should be worked on. All we said is that if there's a plan specification, there's the correlated process, and therefore somewhere there's a realizable, and that the realizable follows from the plan specification. That's very little.

Regarding intention, I view deliberating and arriving at an intention to be a different kind of process. It's certainly a part of what happens when people do things. Perhaps intention is a trigger of sorts. We just didn't talk about it given that it wasn't important for the use we intended. Open world and all that. I know that, personally, I try to stay away from any ontology of cognition because I think in too many cases it's just made up without a strong scientific understanding of cognition. Some day we'll know better, but so far not.

Regarding the simple computer example, and the brain maybe not working that way, the problem with the argument isn't related to the brain. The example doesn't work even in consideration of the reality of how computers work. What you are doing, by saying the disposition is to print strings, or in the example of the gate, is choosing a level of abstraction that *you* are comfortable with. You are comfortable talking at the level of 1s and 0s, even though the next level down is something about charge and current, and the next level down from that is about ions/particles and the  structure of materials. The idea that the plan specification is input to the planned process doesn't give enough - there's no differentiation between a plan specification and information necessary for a process.

You don't change your language to talk about voltage, or coulombs, in order to describe what a computer does, or the relationship between programs and executions. The choice of the primitive being printing some string is completely arbitrary. Computers don't natively print strings. There are a bunch of levels of abstraction before you get to what you seem to want to talk about - the physical dispositions of the material in which the concretizations inhere.

I think it's a reasonable question about whether the appropriate realizable is a disposition. There's a good argument, I think, that it's a role in that it's the sort of thing that's clearly dependent the human/social activity. I think I rejected the idea of role because dispositions had less baggage. Even though there's no axioms to the effect, there's an implicit sense in which roles are conferred, and I thought it more neutral that it be a disposition. If I was doing it over I might simply say the plan specification is concretized as some realizable.

I'd be much happier if the basic notions around realizables were better developed. There are real problems for which we have no guidance. I think I alluded to this in a earlier message. But for now, they are lightly specified, and use is at many levels of abstraction.

Frankly, I don't really understand what the resistance to the notion is about. At the simplest level we're saying that there's a connection between a plan specification and a possible future process. The way we talk about such connections is via realizables, and that's what's done here. Realizables inhere in ICs. A plan specification isn't an IC, so we need another connection and that is concretization. Concretization is similarly loosely specified. We've talking in the past about defining GDCs in terms of the possibility of sharing, but for now that's not how they are defined. I suspect, but I don't have a concrete example at the moment, that there are things we would agree are GDCs some of whose concretizations  can't be shared. An example might be entities that are destroyed when they are 'read'.

I'd be interested in understanding exactly what problems changing this would solve. It seems to me to be the simplest thing to say that's consistent with, and embedded in, BFO, and it's not obviously wrong. I can't imagine that the problems with it are anywhere near the importance of filling in the details that we deliberately omitted, and I don't see that it blocks any work that would further elucidate the subject.

Alan

On Fri, Jan 24, 2020 at 11:26 AM Bill Duncan <wddu...@gmail.com> wrote:

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAM2Eo-doFmG1J9%3Dh1-JRduJPjdwCv1vwTM9WEpBrc4kSL1OcdA%40mail.gmail.com.

[Bill Duncan]

Alan,

Yes. My examples of computer model were simple. If the notion of realizables concertizing plan specifications does hold up well in simple models, then I am skeptical about it holding up in more complex models. If want to explore this in terms of voltages and charges, then please do. It could be quite illuminating :)

I believe I've already expressed my concerns in previous emails.f

Thanks for the feedback!

Bill

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAFKQJ8mNNhtPDoVurpBFndMEacsBq_irhQALTme%2BHv6xsOAUEg%40mail.gmail.com.

[Adrien Barton]

Hi Alan,

One reason I see for not wanting plan specifications (or action specifications) to be concretized by realizable entities is illustrated by the following example.

Consider a stop sign at a crossroad that reads “STOP”. Let’s call ICE0 = ‘STOP’ the information content entity that is concretized by a quality (the white printed pattern “STOP”) inhering in this sign (I suppose that ICE0 is an action specification, and maybe also a plan specification).

Max arrives in his car in front of the stop sign. He reads the sign but for some reason does not intend to follow the instruction to stop. Since he has read the sign, and despite him having no intention (or plan) to follow the instruction to stop, ICE0 is concretized in his cognitive system by an information quality entity IQE1, which is the representation of ICE0 in Max’s cognitive system.

Later, Sam arrives in her car in front of the stop sign. She reads the sign and intends to follow the instruction to stop. Claiming that ICE0 is concretized by a realizable inhering in Sam that would be realized by her stopping would seem to be missing a part of the picture, namely that Sam represents ICE0 in her cognitive system (like Max) independently of whether she intends to follow it or not. Therefore, it may seem more adequate to claim that ICE0 is concretized by a quality IQE2 inhering in Sam’s cognitive system (a representation of ICE0). IQE2 is then used as input of a cognitive process (a decision process) that has as output Sam’s realizable to stop (a plan or an intention, which I believe could indeed be construed as a disposition). 

That being said, you raised a good point, that also applies to decision processes:

>>The idea that the plan specification is input to the planned process doesn't give enough - there's no differentiation between a plan specification and information necessary for a process.

Maybe indeed cognitive representations of plan specifications belong to a specific kind of input into decision processes, that should be analyzed further in order to be distinguished from mere informational input. But to repeat the point I made above, construing the relation between a plan specification and an intention (or plan) to act in a compliant way as a relation of concretization would miss the fact that there needs to be a cognitive representation of a plan specification before a decision to act is taken.

Best,

Adrien

>>> To unsubscribe from this group and stop receiving emails from it, send an email to bfo-d...@googlegroups.com.

>>> To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAM2Eo-crNG_L8XKQKak2XP%3DkdTwLF4KEpGWHUQZD_Ck%3Difx0fA%40mail.gmail.com.
>>>
>>> --
>>> You received this message because you are subscribed to the Google Groups "BFO Discuss" group.

>>> To unsubscribe from this group and stop receiving emails from it, send an email to bfo-d...@googlegroups.com.

>>> To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/E35A775A-89EC-4332-983C-B6FD618CECD3%40infomed.dia.fi.upm.es.
>>
>> --
>> You received this message because you are subscribed to the Google Groups "BFO Discuss" group.

>> To unsubscribe from this group and stop receiving emails from it, send an email to bfo-d...@googlegroups.com.

>> To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAN_82SSn%2BipXnBvgvxz-sYq9MrLPbQ-wfCQDs7u8uats_--%2BBg%40mail.gmail.com.

--
You received this message because you are subscribed to the Google Groups "BFO Discuss" group.

To unsubscribe from this group and stop receiving emails from it, send an email to bfo-d...@googlegroups.com.

[Adrien Barton]

Hi Alan,

One reason I see for not wanting plan specifications (or action specifications) to be concretized by realizable entities is illustrated by the following example.

Consider a stop sign at a crossroad that reads “STOP”. Let’s call ICE0 = ‘STOP’ the information content entity that is concretized by a quality (the white printed pattern “STOP”) inhering in this sign (I suppose that ICE0 is an action specification, and maybe also a plan specification).

Max arrives in his car in front of the stop sign. He reads the sign but for some reason does not intend to follow the instruction to stop. Since he has read the sign, and despite him having no intention (or plan) to follow the instruction to stop, ICE0 is concretized in his cognitive system by an information quality entity IQE1, which is the representation of ICE0 in Max’s cognitive system.

Later, Sam arrives in her car in front of the stop sign. She reads the sign and intends to follow the instruction to stop. Claiming that ICE0 is concretized by a realizable inhering in Sam that would be realized by her stopping would seem to be missing a part of the picture, namely that Sam represents ICE0 in her cognitive system (like Max) independently of whether she intends to follow it or not. Therefore, it may seem more adequate to claim that ICE0 is concretized by a quality IQE2 inhering in Sam’s cognitive system (a representation of ICE0). IQE2 is then used as input of a cognitive process (a decision process) that has as output Sam’s realizable to stop (a plan or an intention, which I believe could indeed be construed as a disposition). 

That being said, you raised a good point, that also applies to decision processes:

>>The idea that the plan specification is input to the planned process doesn't give enough - there's no differentiation between a plan specification and information necessary for a process.

Maybe indeed cognitive representations of plan specifications belong to a specific kind of input into decision processes, that should be analyzed further in order to be distinguished from mere informational input. But to repeat the point I made above, construing the relation between a plan specification and an intention (or plan) to act in a compliant way as a relation of concretization would miss the fact that there needs to be a cognitive representation of a plan specification before a decision to act is taken.

Best,

Adrien

[James A. Overton]

Hi Adrien,

I'm not persuaded that a STOP sign is a plan specification or action specification. As you describe the case, the STOP sign is a trigger for a realizable that Max and Sam bear as trained car drivers, or perhaps as English language speakers.

My understanding is that realizable has a physical basis, and physical triggers. Something causes the brittle teacup to shatter. But we don't seek to specify all the possible triggering causes for shattering. I don't want to try to describe the complex interplay of information flowing through the brain of a lab technician as she performs an experiment. I want to point to the protocol document that she is supposed to be following.

Practically speaking, Bill's proposal is a major change. Bill may have been advocating for this change for a while, but in the decade that I've been on the weekly OBI calls, I've never heard this proposal. Systems that query for specified inputs do no expect to see plan specifications. I would expect this change to break most of the queries for planned processes across the systems I've seen.

I agree with Alan: The current modelling may not be highly detailed, but it's not broken, and I don't see what important problem the proposed change fixes.

James

On Jan 25, 2020, at 09:02, Adrien Barton <adrien...@gmail.com> wrote:



--
You received this message because you are subscribed to the Google Groups "BFO Discuss" group.

To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAPdpnCQVfdoORPM_nreHWvQdPoNYgFP7eb%2B91KFdfM-y7axC6w%40mail.gmail.com.

[Bill Duncan]

James:

As long as good versioning is in place, I doubt that the change would break a lot of systems. I am also skeptical that many OBI users actually make use of the anonymous class 'realizes some (concretizes some plan specification)' for queries. Besides users would have to query for  'has_specified_input some (concretizes some plan specification)', not 'has_specified_input some plan specification'. In any case, OBI's representation will not be consistent with the BFO-ISO definition of 'concretizes'.

You also mentioned processes concretizing GDCs. I think this has place for certain kinds of representations, such as Morse code transmissions. In such cases, there is copying/mapping process between the content of a GDC and the temporal pattern. However, in cases in which an agent (computer or human) executes a plan specification, calling the actions of the agent a pattern that has a "mapping" the GDC seems to stretch the notion of pattern beyond clear understanding.

One thing that confuses about the current OBI representation is how one is represent plan processes that fail.

Realizables are "sure fire" entities. The glass either shatters or it doesn't. But what does it mean to fail in OBI's case. The 'concretization of the plan specification' was either realized or it wasn't. If I am running an experiment and the equipment breaks halfway through the process, how does OBI represent this?

Let,

ps1: the particular plan specification for my experiment

rps1: the particular realizable entity that is the concretization of ps1 and inheres in me

pp1: the particular planned process that realizes rsps1

When I start the experiment, has rps1 been realized? I would say no. There are still crucial parts of experiment that still need to be carried out. You "could" say that rsp1 was realized in some cognitive process, but it is a different conversation. We are considering the planned process being executed by me.

If the experiment successfully finishes, has rsp1 been realized? Under the OBI's model, you can answer yes; rsp1 has been fully "manifested". This seems to be the only case OBI is considering.

But what about cases of equipment malfunctions or other mishaps? The plan specification was still being followed (i.e., it was a specified input the planned process), but it wasn't fully "manifested".

The current theory of realizable entities doesn't allow for partial realizations. My account allows you to represent such failures; ps1 is a specified input to pp1. The full completion of pp1 is a different matter. It important to be able to track failure. The OBI model doesn't permit this.

Thanks for the comments!

Bill

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CDD83340-01E9-4FE3-A9ED-3889B195F768%40overton.ca.

[James A. Overton]

Hi Bill,

My point is that existing systems do no expect the concretization of a plan specification among the specified inputs to a planned process, and that querying for the specified inputs is important for many systems. Having to handle a new kind of specified input for any planned process is a major change. I still haven't heard of any corresponding major benefit.

Your claim that "OBI's representation will not be consistent with the BFO-ISO definition of 'concretizes'" rests on your assertion that "realizable entities are not observable (only their manifestations are observable)". That's an exceedingly narrow notion of "observable" with a large body of philosophy of science literature weighing against it. Realizables do have a physical basis, and we can often observe at least parts of that basis. That assertion ignores Alan's crucial point about levels of abstraction -- another large and fascinating topic with plenty of literature. Your claims about "sure fire" also ignore levels of abstraction. If a glass falls on a carpet and does not shatter, it does not follow that the glass is not brittle. We usually don't want to specify all the micro-conditions under which brittleness can be realized.

Since the venue is the BFO mailing list, I thought this was a discussion about BFO. If this is really a discussion about OBI, it should take place on the OBI tracker, and engage the community of OBI users and developers who depend on this modelling. Failed and incomplete planned processes have been recently discussed by OBI developers and on OBO operations calls.

In short: I agree with Alan the current modelling is adequate; I'm not sure that this proposal works or that it is any better; I don't see what problem the proposal solves; the cost of change is high, and thus the burden of proof is high, and it falls on those proposing change.

I don't think that I have anything more to say about this topic.

James

On Jan 25, 2020, at 10:30, Bill Duncan <wddu...@gmail.com> wrote:



To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAM2Eo-dLQRZ4ih2FtFPhcdOn%2Bryw37%3DKC0PJM5f26Eh4%2BsX4HQ%40mail.gmail.com.

[Bill Duncan]

James:

I also posted my original discussion on the obo mailing list, no one responded, so I assumed no one was interested. The major parties in OBI (e.g,, Bjorn, Chris, you, Rand, etc.) are also on this list. I don't see a need for duplication. You are welcome to forward the conversation if you like.

You wanted a use case. I gave you one involving failure of planned processes. It would be helpful if you could give such an account using the current OBI framework. This, I think, is a failing of OBI.

Your claim about levels of abstraction is misleading. I asked Alan to illuminate such an example. I have yet to hear one. Perhaps you can provide one that shows how my more perspicacious account fails.

Invoking the history of philosophy, is not a convincing argument. BFO takes into account some things from this literature, and rejects others. Sorry, but you need to be more precise. I have spent a lot of key strokes attempting to provide examples that every day ontologist can understand. Not hand waiving notions about levels of abstraction, complex relationships amongst materials entities, triggers and the like. If you think that such notions are convincing against what I am proposing, I only ask that you provide a concrete example of one.

The glass dropping on the floor and not shattering is just an example of the disposition not being realized (in BFO terms). Can you please explain how this related the example of a failing planned process that I provided. Do you have a theory of partial realizations that you wish to propose?

Thanks for the feedback!

Bill

P.S. I use large font b/c my eyesight is succumbing to the ravages of age :(

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/18E84A62-7B9D-4A71-A71C-3BECF539FDB6%40overton.ca.

[Alan Ruttenberg]

On Saturday, January 25, 2020, Adrien Barton <adrien...@gmail.com> wrote:
> Hi Alan,
>
> One reason I see for not wanting plan specifications (or action specifications) to be concretized by realizable entities is illustrated by the following example.
>
> Consider a stop sign at a crossroad that reads “STOP”. Let’s call ICE0 = ‘STOP’ the information content entity that is concretized by a quality (the white printed pattern “STOP”) inhering in this sign (I suppose that ICE0 is an action specification, and maybe also a plan specification).
>
> Max arrives in his car in front of the stop sign. He reads the sign but for some reason does not intend to follow the instruction to stop. Since he has read the sign, and despite him having no intention (or plan)

Here it seems you are equating having an intention with having a plan. While I recognize that this equivalence is recognized in some uses of the word plan, the word can also have a more neutral interpretation and this is the one I intend when I use it in these conversations.

There is a term plan in OBI which does include the idea that there is a commitment to execute, however we don't say that a plan specification is concretized as an obi:plan. The intended reading is just that the plan spec is concretized as some disposition that's realized as a planned process. A process, of course, may realize multiple dispositions.

However, let's work with the example. My sense of what is happening is that first the stop sign is comprehended and predisposes the driver to stop. The disposition may or may not be realized. Consider a situation in which the driver first decides not to stop, but then sees police at the intersection. That decision might be revoked.  To my mind the choices and deliberation happen *after* the sign is comprehended / concretized, and they seem to act as trigger/blocker for the concretized stop disposition.

to follow the instruction to stop, ICE0 is concretized in his cognitive system by an information quality entity IQE1, which is the representation of ICE0 in Max’s cognitive system.

As I think I've mentioned, this step of assuming that the representation of in the brain is a quality is not obvious and is not backed by science that I'm aware of. There's certainly some specifically dependent continuant involved. Its the choice of which one that is harder to assert.

>
> Later, Sam arrives in her car in front of the stop sign. She reads the sign and intends to follow the instruction to stop. Claiming that ICE0 is concretized by a realizable inhering in Sam that would be realized by her stopping would seem to be missing a part of the picture, namely that Sam represents ICE0 in her cognitive system (like Max) independently of whether she intends to follow it or not.

It does not follow that the independence means that the representation isn't a disposition. Unless you only believe the disposition can exist if there's intent. I don't. The dispositions of inanimate objects exist just fine without intent, and I see no reason to insist that dispositions in humans require it. I'm pretty sure that's not the case for roles. You can have a role before there's any situation in which the realization could be triggered.

So 2 independent things, at least, when a plan is intended to be followed. The disposition, and the deliberative process leading to intent.

Moreover it's not clear to me that in executing planned processes there is always 'intent'. First, is easy to talk loosely when there's been no offer of any definition of intent. Second you may plan to, e.g. fight back in a certain way when attacked, but when the attack happens you may react with muscle memory/ 'instict' without any deliberative process. Again this depends on what one defines intent to be.

Therefore, it may seem more adequate to claim that ICE0 is concretized by a quality IQE2 inhering in Sam’s cognitive system (a representation of ICE0).

If you were certain that all such representions are qualities. I'm. And I don't think it's obvious.

IQE2 is then used as input of a cognitive process (a decision process) that has as output Sam’s realizable to stop (a plan or an intention, which I believe could indeed be construed as a disposition). 

Do you have an argument ruling out that the deliberation and intent function as trigger to an existing disposition as opposed to *being* the disposition, as I understand your analysis to be saying.

> That being said, you raised a good point, that also applies to decision processes:
>>>The idea that the plan specification is input to the planned process doesn't give enough - there's no differentiation between a plan specification and information necessary for a process.
> Maybe indeed cognitive representations of plan specifications belong to a specific kind of input into decision processes, that should be analyzed further in order to be distinguished from mere informational input. But to repeat the point I made above, construing the relation between a plan specification and an intention (or plan) to act in a compliant way as a relation of concretization would miss the fact that there needs to be a cognitive representation of a plan specification before a decision to act is taken.

A 'cognitive representation'. What's that? I'm all for expanding on this notion, if possible. So far I haven't heard much defining the nature of these representations and how they function. To me it looks like the model is that our decision making looks like what a simple computer does - there are some symbolic inputs processed in a data transform of sorts. I think the idea that all cognition takes that form has been debuncked a ways back when ai planning found it didn't work. The proposed alternative to that, at the time, was 'situated action'

But let's continue talking...

Regards,
Alan
(Disclaimer: writing on the phone while waiting for a plane to take off, so hopefully my response is coherent. )

> To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/9482a134-8917-4af6-a6df-8154e9cbb613%40googlegroups.com.
>

[Alan Ruttenberg]

I think that you missed my point about your example. I wasn't trying to say that your dispositional account was wrong. Rather, I was saying that idea that it was the only correct account wasn't convincing. It seems to me there are two possible approaches to answering this question. Either there's a 'right' answer that exists at the right granularity, or we accept that dispositon talk can be had at many different levels. I think you are going for the 'single right level' and my response to that is to point out how the choice you made wasn't satisfactory as the one answer. Rather, current usage as well as practicality suggests to me instead that there are multiple levels of granularity at which disposition talk in the spirit of BFO is valid.

Best,
Alan

> To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAM2Eo-e6Hm%2Bfg5rEKicV%3DBg07JtvDmWKn%2BO9fNy8wEUeBCTn%3Dg%40mail.gmail.com.
>

[Adrien Barton]

[James wrote:]

> I'm not persuaded that a STOP sign is a plan specification or action specification. As 

> you describe the case, the STOP sign is a trigger for a realizable that Max and Sam 

> bear as trained car drivers, or perhaps as English language speakers.

In BFO, realizables are triggered by occurrents. The STOP sign, however, is a continuant. Hence, it cannot be a trigger. Maybe the cognitive process of seeing the STOP sign could be a trigger though. And indeed, I agree that Max and Sam, as trained car drivers, have a general disposition to stop when seeing a STOP sign. But it does not mean that those are the only dispositions relevant here (more on this below); and moreover, this scenario is complexified by the fact that a deontic is involved, namely the obligation to stop at a STOP sign.

So if you prefer, consider the simpler example of a part of a recipe reading ICE3 = ‘crack an egg’ (I’m not sure why you have doubts that ICE0=‘STOP’ would be an action specification, but I hope it’s less controversial that ICE3 is an action specification?). People don’t necessarily develop a disposition to crack an egg when they read ICE3. However, if Sam decides to follow the recipe, she will develop a disposition (which I call her intention to crack an egg) that will be realized by her cracking an egg, in case this disposition is not blocked. This disposition may appear some time later after she reads ICE3 though (for example, she might deliberate about whether she will follow the recipe), that’s why I don’t think it can be the concretization of ICE3.

[Alan wrote:]

> Here it seems you are equating having an intention with having a plan. While I 

> recognize that this equivalence is recognized in some uses of the word plan, the 

> word can also have a more neutral interpretation and this is the one I intend when 

> I use it in these conversations.

 

Actually I was trying not to commit too much about this, as this seems to me to be a secondary question here, that’s why I mentioned very generally “an intention or a plan” without discussing it. I won’t use the word “plan” anymore here.

> However, let's work with the example. My sense of what is happening is that first 

> the stop sign is comprehended and predisposes the driver to stop. The disposition 

> may or may not be realized. Consider a situation in which the driver first decides 

> not to stop, but then sees police at the intersection. That decision might be 

> revoked.  To my mind the choices and deliberation happen *after* the sign is 

> comprehended / concretized, and they seem to act as trigger/blocker for the 

> concretized stop disposition.

I would agree that the deliberation happens after the ICE from the sign (ICE0) is concretized in her cognitive system, but I don’t think that ICE0 is concretized by a disposition of the driver to stop.

Let’s consider the above example of ICE3 = ‘crack an egg’. Suppose that Max reads ICE3 but has no intention to follow the instruction: then it does not seem that ICE3 is concretized by a disposition of Max to crack an egg.

> this step of assuming that the representation of in the 

> brain is a quality is not obvious and is not backed by science that I'm aware of. 

> There's certainly some specifically dependent continuant involved. Its the choice of 

> which one that is harder to assert.

I agree that this question is complicated. However, my main point here is that a concretization of an action specification such as ‘STOP’ (resp. ‘crack an egg’) in a reader is not a disposition to stop (resp. a disposition to crack an egg).

> It does not follow that the independence means that the representation isn't a 

> disposition. Unless you only believe the disposition can exist if there's intent. I 

> don't. The dispositions of inanimate objects exist just fine without intent, and I see 

> no reason to insist that dispositions in humans require it. I'm pretty sure that's not 

> the case for roles. You can have a role before there's any situation in which the 

> realization could be triggered.

I certainly agree that dispositions can exist without intent – as you mention, inanimate objects do have plenty of dispositions. For example, Sam has a disposition d0 to push the break pedal when some heavy object pushes down her right foot on the break pedal. That’s purely mechanical, and does not imply any intention. But this disposition d0 of Sam has nothing to do with her reading ICE0=’STOP’.

On the other hand, I’d say that reading ICE0 leads to the apparition of a disposition d1 of Sam to push the break pedal intentionally (note that d0 and d1 do not have the same kinds of triggers: d1 is not triggered by a heavy object pushing down Sam’s right foot; hence those are two different dispositions). d1 is related in some way to the concretization of ICE0 in her cognitive system, but I don’t think it is identical to this concretization, for the reason I mentioned above. Rather, I’d say that d1 is the output of Sam’s decision process after ICE0 is concretized in her cognitive system.

> Moreover it's not clear to me that in executing planned processes there is always 

> 'intent'. First, is easy to talk loosely when there's been no offer of any definition of 

> intent. Second you may plan to, e.g. fight back in a certain way when attacked, but 

> when the attack happens you may react with muscle memory/ 'instinct' without 

> any deliberative process. Again this depends on what one defines intent to be.

I was using the word “intention” rather than “intent”, as the former does not seem to imply that there has been a deliberative process. In particular, an intention (as I understand it) can result not only from a deliberative decision process, but also from a heuristic decision process. I can have a disposition to act that is due to an instinctive decision process, but this disposition might still be blocked (although such dispositions may be harder to block that the dispositions to act that result from a deliberative decision process).

> Do you have an argument ruling out that the deliberation and intent function as 

> trigger to an existing disposition as opposed to *being* the disposition, as I 

> understand your analysis to be saying.

It seems to me that intentions behave more like continuants than like occurrents: my intention to read this book can exist fully at both t1 and t2, and wax and wane as time passes.

Moreover, intentions seem to have a dispositional character, in the sense that my intention to do A can lead to me doing A, or can be blocked. For example, I formed the intention to go for lunch but suddenly I remember that I have to write this email, so my intention is not realized. Or more radically, I decided to stand up but suddenly I’m paralyzed, so I can’t.

> A 'cognitive representation'. What's that? I'm all for expanding on this notion, if 

> possible. So far I haven't heard much defining the nature of these representations 

> and how they function. To me it looks like the model is that our decision making 

> looks like what a simple computer does - there are some symbolic inputs 

> processed in a data transform of sorts. I think the idea that all cognition takes that 

> form has been debunked a ways back when ai planning found it didn't work. The 

> proposed alternative to that, at the time, was 'situated action'

I certainly agree that much more remains to be done in the field of the ontology of mental entities. Here, I was just pointing to one direction in which a (maybe naïve) ontology of mental entities could be developed to account for some phenomena. But we have to get progressively closer to the underlying cognitive science indeed.

To summarize, let me rephrase my point more carefully by distinguishing two questions:

Q1) Is an ICE always concretized by a quality? 

This is admittedly a complex question, in particular because the links between qualities and realizable entities are complex. I won’t argue about this here.

Q2) Is an action specification that specifies to do A (such as ICE0=‘STOP’ or ICE3=‘crack an egg’) concretized in a reader by a disposition to do A?

The answer, I think, is “no”: the disposition to do A intentionally will only appear after the action specification was concretized in the agent’s cognitive system and a (deliberative or heuristic) decision process has taken place (and the dispositions of the agent to do A non-intentionally, in a purely mechanical way, have nothing to do with this action specification, so they cannot be the concretization of this action specification either).

Best,

Adrien

[Alan Ruttenberg]

On Tue, Jan 28, 2020 at 8:08 PM Adrien Barton <adrien...@gmail.com> wrote:

[James wrote:]

> I'm not persuaded that a STOP sign is a plan specification or action specification. As 

> you describe the case, the STOP sign is a trigger for a realizable that Max and Sam 

> bear as trained car drivers, or perhaps as English language speakers.

In BFO, realizables are triggered by occurrents.

While that is compatible with BFO, BFO says no such thing. Many discussions of dispositions talk about circumstances in which the disposition manifests. Personally, I'm not convinced by the trigger=process argument. Consider the disposition of a piece of paper to catch fire at a certain temperature, or at a temperature that depends on the humidity. The important thing here is that there's a quality threshold. Of course, you can coerce any continuant into a process - the process during which it comes to be, but that seems to me to be secondary. There are many processes that can raise the temperature. What they all have in common is not the manner in which they do that, but that the type of the temperature / humidity of the surrounding air reaches the threshold.  Circumstances, BTW, sound to me like states of affairs, or what Werner calls configurations. I think BFO needs to evolve to be able to talk about such.

 

The STOP sign, however, is a continuant. Hence, it cannot be a trigger. Maybe the cognitive process of seeing the STOP sign could be a trigger though. And indeed, I agree that Max and Sam, as trained car drivers, have a general disposition to stop when seeing a STOP sign. But it does not mean that those are the only dispositions relevant here (more on this below); and moreover, this scenario is complexified by the fact that a deontic is involved, namely the obligation to stop at a STOP sign.

So if you prefer, consider the simpler example of a part of a recipe reading ICE3 = ‘crack an egg’ (I’m not sure why you have doubts that ICE0=‘STOP’ would be an action specification, but I hope it’s less controversial that ICE3 is an action specification?).

It's very hard to disambiguate whether part of a plan specification is a goal specification versus an objective specification. For certain kinds of machines the distinction is reasonably clear - if the machine has or needs to feedback to realize the process then it is an objective specification. For instance: release a latch, resulting in a hatch opening - no feedback used or necessary. For humans it's more difficult. Arguing against a stop sign being an action specification is that the action achieving the goal of stopping can be rather varied. If on a bike you might apply hand brakes, or you might slow down and use your feet for friction. In a car you might press the brake, or, if it fails you might use the hand brake. If you are walking then you stop walking.  All these are different actions.

People don’t necessarily develop a disposition to crack an egg when they read ICE3.

Well, you are asserting this. However that's the crux of the argument, so it isn't resolved with an assertion one way or the other.

 

However, if Sam decides to follow the recipe, she will develop a disposition (which I call her intention to crack an egg) that will be realized by her cracking an egg, in case this disposition is not blocked. This disposition may appear some time later after she reads ICE3 though (for example, she might deliberate about whether she will follow the recipe), that’s why I don’t think it can be the concretization of ICE3.

I think we're probably going in circles. I've tried to distinguish two different things: deliberation leading to intention as opposed to the recipe concretized as disposition. 'Deliberation' needs to be unpacked. I'm taking it to be a conscious process. In that sense, it seems to me that the disposition might be realized with or without deliberation.

[Alan wrote:]

> Here it seems you are equating having an intention with having a plan. While I 

> recognize that this equivalence is recognized in some uses of the word plan, the 

> word can also have a more neutral interpretation and this is the one I intend when 

> I use it in these conversations.

 

Actually I was trying not to commit too much about this, as this seems to me to be a secondary question here, that’s why I mentioned very generally “an intention or a plan” without discussing it. I won’t use the word “plan” anymore here.

> However, let's work with the example. My sense of what is happening is that first 

> the stop sign is comprehended and predisposes the driver to stop. The disposition 

> may or may not be realized. Consider a situation in which the driver first decides 

> not to stop, but then sees police at the intersection. That decision might be 

> revoked.  To my mind the choices and deliberation happen *after* the sign is 

> comprehended / concretized, and they seem to act as trigger/blocker for the 

> concretized stop disposition.

I would agree that the deliberation happens after the ICE from the sign (ICE0) is concretized in her cognitive system, but I don’t think that ICE0 is concretized by a disposition of the driver to stop.

Yes. We think differently about this :-)

 

Let’s consider the above example of ICE3 = ‘crack an egg’. Suppose that Max reads ICE3 but has no intention to follow the instruction: then it does not seem that ICE3 is concretized by a disposition of Max to crack an egg.

 

They may have no intention now. That doesn't mean they won't have an intention later. What I'm arguing is that the ability, a disposition, is planted when reading the instructions. Ability is a little too strong a word here - I'll try to think of another. Maybe disposition ;-)

> this step of assuming that the representation of in the 

> brain is a quality is not obvious and is not backed by science that I'm aware of. 

> There's certainly some specifically dependent continuant involved. Its the choice of 

> which one that is harder to assert.

I agree that this question is complicated. However, my main point here is that a concretization of an action specification such as ‘STOP’ (resp. ‘crack an egg’) in a reader is not a disposition to stop (resp. a disposition to crack an egg).

Again an assertion.

 

> It does not follow that the independence means that the representation isn't a 

> disposition. Unless you only believe the disposition can exist if there's intent. I 

> don't. The dispositions of inanimate objects exist just fine without intent, and I see 

> no reason to insist that dispositions in humans require it. I'm pretty sure that's not 

> the case for roles. You can have a role before there's any situation in which the 

> realization could be triggered.

I certainly agree that dispositions can exist without intent – as you mention, inanimate objects do have plenty of dispositions. For example, Sam has a disposition d0 to push the break pedal when some heavy object pushes down her right foot on the break pedal. That’s purely mechanical, and does not imply any intention. But this disposition d0 of Sam has nothing to do with her reading ICE0=’STOP’.

The 'purely mechanical' part is interesting. But for me the 'purely mechanical' is the distinction between whether the realization of an action specification or a goal specification. Since plan specifications can include both, I don't think this is relevant to the determination of whether there's a disposition or not.

 

On the other hand, I’d say that reading ICE0 leads to the apparition of a disposition d1 of Sam to push the break pedal intentionally (note that d0 and d1 do not have the same kinds of triggers: d1 is not triggered by a heavy object pushing down Sam’s right foot; hence those are two different dispositions).

This is either overstepping or trivializing intention. If anything a person does manifests intention then all 'intention' adds is that the participant is a person. Also consider 'muscle memory'. There are actions that as learned and practiced become 'automatic'. As an early piano player there is a lot of conscious thought involved in the actions of hitting the keys. However when practiced there is relatively little conscious deliberation between reading the notes in a music score and playing them.  

 

d1 is related in some way to the concretization of ICE0 in her cognitive system, but I don’t think it is identical to this concretization, for the reason I mentioned above. Rather, I’d say that d1 is the output of Sam’s decision process after ICE0 is concretized in her cognitive system.

> Moreover it's not clear to me that in executing planned processes there is always 

> 'intent'. First, is easy to talk loosely when there's been no offer of any definition of 

> intent. Second you may plan to, e.g. fight back in a certain way when attacked, but 

> when the attack happens you may react with muscle memory/ 'instinct' without 

> any deliberative process. Again this depends on what one defines intent to be.

I was using the word “intention” rather than “intent”, as the former does not seem to imply that there has been a deliberative process. In particular, an intention (as I understand it) can result not only from a deliberative decision process, but also from a heuristic decision process. I can have a disposition to act that is due to an instinctive decision process, but this disposition might still be blocked (although such dispositions may be harder to block that the dispositions to act that result from a deliberative decision process).

Good. We understand intent(ion) in the same way. This would seem to strengthen my argument, since it's clear that the same instructions can either be acted on with conscious deliberation, or not.

 

> Do you have an argument ruling out that the deliberation and intent function as 

> trigger to an existing disposition as opposed to *being* the disposition, as I 

> understand your analysis to be saying.

It seems to me that intentions behave more like continuants than like occurrents: my intention to read this book can exist fully at both t1 and t2, and wax and wane as time passes.

Ok.  But as I've said, I don't subscribe to the 'trigger must be an occurrent' perspective.

Moreover, intentions seem to have a dispositional character, in the sense that my intention to do A can lead to me doing A, or can be blocked. For example, I formed the intention to go for lunch but suddenly I remember that I have to write this email, so my intention is not realized. Or more radically, I decided to stand up but suddenly I’m paralyzed, so I can’t.

This seems reasonable.

 

> A 'cognitive representation'. What's that? I'm all for expanding on this notion, if 

> possible. So far I haven't heard much defining the nature of these representations 

> and how they function. To me it looks like the model is that our decision making 

> looks like what a simple computer does - there are some symbolic inputs 

> processed in a data transform of sorts. I think the idea that all cognition takes that 

> form has been debunked a ways back when ai planning found it didn't work. The 

> proposed alternative to that, at the time, was 'situated action'

I certainly agree that much more remains to be done in the field of the ontology of mental entities. Here, I was just pointing to one direction in which a (maybe naïve) ontology of mental entities could be developed to account for some phenomena. But we have to get progressively closer to the underlying cognitive science indeed.

To summarize, let me rephrase my point more carefully by distinguishing two questions:

Q1) Is an ICE always concretized by a quality? 

This is admittedly a complex question, in particular because the links between qualities and realizable entities are complex. I won’t argue about this here.

Q2) Is an action specification that specifies to do A (such as ICE0=‘STOP’ or ICE3=‘crack an egg’) concretized in a reader by a disposition to do A?

The answer, I think, is “no”: the disposition to do A intentionally will only appear after the action specification was concretized in the agent’s cognitive system and a (deliberative or heuristic) decision process has taken place (and the dispositions of the agent to do A non-intentionally, in a purely mechanical way, have nothing to do with this action specification, so they cannot be the concretization of this action specification either).

I understand your position. I think, at best, I might be convinced, at this point, that the ICE isn't always concretized as a disposition But rather than being convinced, I'd prefer we had better theories that could resolve this sort of question without us having to resort to opinions.

A point of clarification. By suggesting that there's a disposition that concretizes the plan specification I don't mean to rule out that there are other concretizations involved. It seems reasonable to me that in some cases a set of things can happen. You might remember the instructions to the level of quoting the wording, and whatever that memory is, it seems like a different concretization than the one I see as manifesting in action.

Regards,

Alan 

 

Best,

Adrien

--
You received this message because you are subscribed to the Google Groups "BFO Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bfo-discuss...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/473a26ed-ecb5-43e4-aa73-0c111d7d5398%40googlegroups.com.

[Alan Ruttenberg]

A subsequent thought that amused me: The argument that dispositions always have occurrent triggers brings to mind a phrase I internalized when raising my children: It's not the destination, it's the journey. A manifestation of this view: It's not whether you win or lose, but how you play the game. But sometimes it's not the journey, it's the destination, the achievement  that triggers a process - live receiving the prize money in a race. 

Following up on the 'circumstances' vs 'occurrent' trigger point, Here's what the ISO BFO elucidation of disposition says:

ELUCIDATION: b is a disposition means:
b is a realizable entity
& b's bearer is some material entity
& b is such that if it ceases to exist, then its bearer is physically changed,
& b's realization occurs when and because this bearer is in some special physical circumstances,
& this realization occurs in virtue of the bearer's physical make-up [062-ISO]

Alan

[Bill Duncan]

Hi Alan,

Thanks for the feedback. Some comments:

1. I wasn't saying that my computer example was the only correct account. Rather, I was operating under that assumption that the process of a computer executing a program is a kind of planned process. This seems pretty reasonable to me. If this is the case, then my example shows a case in which the axiom "realizes some (concretizes some plan specification) doesn't hold.

2. I don't think I ever said that ICE's could only be concretized as qualities. I leave open the possibility concretizations may be processes. For things such as the voltages in memory electronics this actually may be more accurate.

3. The range of has_specified_input is continuant. So, the axiom "has_specified_input some (concretizes some plan specification)" still leaves room for such concretizations to be realizables. We need to work out what it means for a realizable to be a specified input though.

4. I have stayed away from inventions and the like. My intention to realized the plan may be realized in some cognitive process, although I may not physically carry out the plan. I thought planned processes were about actual carrying plans, not intentions.

5. Do you have an account of how represent errors/failures if when plan specifications are concretized as realizables?

Thanks for the comments!

Bill

To view this discussion on the web visit https://groups.google.com/d/msgid/bfo-discuss/CAFKQJ8ngfkGA5dmMjda0qT%3Dvk-efkhxQg9V_jepqDyJ0LZ4pPQ%40mail.gmail.com.