Issue 10621: Definition of 'question' (sbvr-ftf) Source: NIST (Mr. Edward J. Barkmeyer, edbark(at)nist.gov) Nature: Uncategorized Issue Severity: Summary: In clause 8.1.3, 'question' is defined as "meaning of an interrogatory". Assuming that "interrogatory" means "interrogatory sentence", this defines the meaning in terms of the form of expression. Because the same interrogatory sentence may have more than one meaning, depending on the context of its utterance, this definition means: the meaning of an interrogatory sentence in context. But the same meaning extends across (some) contexts, while the answer may be different in each such context. That is, some part of the context of utterance affects the interpretation of the utterance; other parts of the context of utterance only affect its result. So this is a poor definition. A question is really an "operation" on a concept or proposition that creates a new form of meaning. One can describe a 'question' as a function applied to a concept or proposition whose result is its extension. There are 3 kinds of questions: - a concept question (What), which asks for the extension of a concept, - a proposition question (Whether), which asks for the truth-value (technically the extension) of a proposition. - a propositional relationship question (Why), which asks for the extension of a 2nd-order proposition about propositions (or objectified 'states of affairs' identified by propositions) -- the set of all propositions P such that 'P causes Q' or 'P entails Q'. Since clause 9 apparently supports 'question' meanings in exactly this way, SBVR should define 'question' to be a meaning that explicitly refers to the extension of a concept. That is, the question means its answer. It should not mean the action of asking. Resolution: The FTF discussed this issue at length and could not agree on a resolution. Revised Text: Actions taken: January 23, 2007: received issue Discussion: End of Annotations:===== te: Tue, 23 Jan 2007 15:13:48 -0500 From: Ed Barkmeyer Reply-To: edbark@nist.gov Organization: NIST User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax) X-Accept-Language: en, fr, de, pdf, it, nl, sv, es, ru To: issues@omg.org Subject: SBVR issue -- definition of 'question' X-MailScanner-Information: Please contact postmaster@mel.nist.gov for more information X-MailScanner: Found to be clean X-MailScanner-SpamCheck: X-MailScanner-From: edbark@nist.gov X-Spam-Status: No Doc: SBVR, dtc/06-03-02 Date: March 2006 Version: Final Adopted Specification Chapter: 8.1.3 Related issues: 9715 Title: Definition of 'question' Source: Ed Barkmeyer, NIST, edbark@nist.gov Description: In clause 8.1.3, 'question' is defined as "meaning of an interrogatory". Assuming that "interrogatory" means "interrogatory sentence", this defines the meaning in terms of the form of expression. Because the same interrogatory sentence may have more than one meaning, depending on the context of its utterance, this definition means: the meaning of an interrogatory sentence in context. But the same meaning extends across (some) contexts, while the answer may be different in each such context. That is, some part of the context of utterance affects the interpretation of the utterance; other parts of the context of utterance only affect its result. So this is a poor definition. A question is really an "operation" on a concept or proposition that creates a new form of meaning. One can describe a 'question' as a function applied to a concept or proposition whose result is its extension. There are 3 kinds of questions: - a concept question (What), which asks for the extension of a concept, - a proposition question (Whether), which asks for the truth-value (technically the extension) of a proposition. - a propositional relationship question (Why), which asks for the extension of a 2nd-order proposition about propositions (or objectified 'states of affairs' identified by propositions) -- the set of all propositions P such that 'P causes Q' or 'P entails Q'. Since clause 9 apparently supports 'question' meanings in exactly this way, SBVR should define 'question' to be a meaning that explicitly refers to the extension of a concept. That is, the question means its answer. It should not mean the action of asking. -- Edward J. Barkmeyer Email: edbark@nist.gov National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Subject: RE: issue 10621 -- SBVR FTF issue Date: Tue, 23 Jan 2007 15:13:13 -0700 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: issue 10621 -- SBVR FTF issue Thread-Index: Acc/NINmvgPt1EszTGWHupzHqa2EeAABfQlw From: "Andy Carver" To: Ed, I have a couple of comments on this issue: Your issue is .assuming that .interrogatory. means .interrogatory sentence... And yet I would think this assumption is mistaken. It is not an interrogatory sentence that is in view, but an interrogatory speech act, in the technical sense intended in .speech act theory.. And the whole problem statement, it seems to me, assumes the former definition of .interrogatory.; because once you accept the latter instead, I don.t see that context affects the meaning. I am fine with your .3 kinds of questions.. But this seems to vitiate your suggested, new definition of question (.a meaning that explicitly refers to the extension of a concept.), since that definition only seems relevant to .What. questions (those which ask .for the extension of a concept.). Am I missing something here? Cheers, Andy -------------------------------------------------------------------------------- From: Juergen Boldt [mailto:juergen@omg.org] Sent: Tuesday, January 23, 2007 2:20 PM To: issues@omg.org; sbvr-ftf@omg.org Subject: issue 10621 -- SBVR FTF issue This is issue # 10621 Source: Ed Barkmeyer, NIST, edbark@nist.gov Definition of 'question' In clause 8.1.3, 'question' is defined as "meaning of an interrogatory". Assuming that "interrogatory" means "interrogatory sentence", this defines the meaning in terms of the form of expression. Because the same interrogatory sentence may have more than one meaning, depending on the context of its utterance, this definition means: the meaning of an interrogatory sentence in context. But the same meaning extends across (some) contexts, while the answer may be different in each such context. That is, some part of the context of utterance affects the interpretation of the utterance; other parts of the context of utterance only affect its result. So this is a poor definition. A question is really an "operation" on a concept or proposition that creates a new form of meaning. One can describe a 'question' as a function applied to a concept or proposition whose result is its extension. There are 3 kinds of questions: - a concept question (What), which asks for the extension of a concept, - a proposition question (Whether), which asks for the truth-value (technically the extension) of a proposition. - a propositional relationship question (Why), which asks for the extension of a 2nd-order proposition about propositions (or objectified 'states of affairs' identified by propositions) -- the set of all propositions P such that 'P causes Q' or 'P entails Q'. Since clause 9 apparently supports 'question' meanings in exactly this way, SBVR should define 'question' to be a meaning that explicitly refers to the extension of a concept. That is, the question means its answer. It should not mean the action of asking. Juergen Boldt Director, Member Services Object Management Group 140 Kendrick St Building A Suite 300 Needham, MA 02494 USA tel: +1 781 444 0404 x 132 fax: +1 781 444 0320 email: juergen@omg.org www.omg.org Date: Wed, 24 Jan 2007 17:29:40 -0500 From: Ed Barkmeyer Reply-To: edbark@nist.gov Organization: NIST User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax) X-Accept-Language: en, fr, de, pdf, it, nl, sv, es, ru To: Andy Carver CC: sbvr-ftf@omg.org Subject: Re: issue 10621 -- SBVR FTF issue X-MailScanner-Information: Please contact postmaster@mel.nist.gov for more information X-MailScanner: Found to be clean X-MailScanner-SpamCheck: X-MailScanner-From: edbark@nist.gov X-Spam-Status: No Andy Carver wrote: * Your issue is "assuming that 'interrogatory' means 'interrogatory sentence'". And yet I would think this assumption is mistaken. It is not an interrogatory sentence that is in view, but an interrogatory speech act, in the technical sense intended in "speech act theory". And the whole problem statement, it seems to me, assumes the former definition of "interrogatory"; because once you accept the latter instead, I don't see that context affects the meaning. Thank you. Neither my ODE nor my Webster defines to "interrogatory" as a term in speech act theory. The essence of both definitions is that the noun 'interrogatory' means a written question that requires a formal response. According to SBVR, therefore, that is the meaning. We apparently need at least a minor editorial change to associate it with speech act theory. At the same time, my assumption that it meant an interrogative sentence was unwarranted. And the arguments about relationship to expression were ignoratio elenchi. I agree. I assert, however, that my statement that a question is a speech act ("operation") involving a concept or proposition must then be what was intended. As a speech act, it differs from an assertion, for example, but not from a "proposition", which (we say) is a meaning, something acted on by a speech act. My main concern was to get the SBVR subtypes of 'meaning' to be related by some principle. And I can't find the principle that puts questions in the same taxonomy with concept and proposition. * I am fine with your "3 kinds of questions". But this seems to vitiate your suggested, new definition of question ("a meaning that explicitly refers to the extension of a concept"), since that definition only seems relevant to "What" questions (those which ask "for the extension of a concept"). Am I missing something here? I only proposed that because it matches Don's 'question formulation' in clause 9, which, according to Don, treats Whether questions as projections whose result space is the powerset of {true, false}. Put simply, I don't know what to do with questions, but the current text of SBVR once again demonstrates Pope's aphorism. Since you do seem to know what the relationship of questions to meanings is, perhaps you could suggest a solution. (And if "no change" is the solution, it resolves the issue.) Thanks, -Ed -- Edward J. Barkmeyer Email: edbark@nist.gov National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." Subject: RE: issue 10621 -- SBVR FTF issue Date: Wed, 24 Jan 2007 17:24:59 -0700 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: issue 10621 -- SBVR FTF issue Thread-Index: AcdABygWwT5QSRYaSM+aFIO6IIQatQADOl9w From: "Andy Carver" To: Cc: X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id l0P0HQiN022495 Ed, You wrote, "My main concern was to get the SBVR subtypes of 'meaning' to be related by some principle. And I can't find the principle that puts questions in the same taxonomy with concept and proposition." I must admit I hadn't noticed that very interesting and important problem. And now that I've pointed out that "question" is a kind of speech act, it makes the problem appear even more acute; because one definitely cannot say this same thing about "proposition" or "concept". So are we mixing apples and oranges, and making a taxonomy error? You've got me wondering too. Speech acts often relate to propositions by having a propositional component, as well as a performative component: for example, the sort of speech act called "assertion" would have a meaning-component, viz. the proposition being asserted as true (also called the "locutionary" component), and a performative component, the act/attitude of asserting it. Likewise would a "true/false" question; but here the performative component would obviously be a different one. "What" questions, as you called them, would NOT have a propositional component, since they're not asking for a truth value but for an extension. But perhaps we could say they still have a MEANING-component, viz. the concept whose extension they're requesting. But I'm not sure whether standard speech-act theory allows for such thin spreading of the idea of the "locutionary meaning"; I'd have to research that. In any case, the interesting thing right now is that speech acts are not the meanings, but rather, they contain such "locutionary" meanings but also "performative" and "perlocutionary" meanings. So at this point, it seems to me you've identified a real quandary in our taxonomy. Wish I could shed more light on it, but right now that's the extent of my wisdom. Cheers, Andy -----Original Message----- From: Ed Barkmeyer [mailto:edbark@nist.gov] Sent: Wednesday, January 24, 2007 3:30 PM To: Andy Carver Cc: sbvr-ftf@omg.org Subject: Re: issue 10621 -- SBVR FTF issue Andy Carver wrote: > * Your issue is "assuming that 'interrogatory' means > 'interrogatory sentence'". And yet I would think this assumption is > mistaken. It is not an interrogatory sentence that is in view, but an > interrogatory speech act, in the technical sense intended in "speech act > theory". And the whole problem statement, it seems to me, assumes the > former definition of "interrogatory"; because once you accept the latter > instead, I don't see that context affects the meaning. Thank you. Neither my ODE nor my Webster defines to "interrogatory" as a term in speech act theory. The essence of both definitions is that the noun 'interrogatory' means a written question that requires a formal response. According to SBVR, therefore, that is the meaning. We apparently need at least a minor editorial change to associate it with speech act theory. At the same time, my assumption that it meant an interrogative sentence was unwarranted. And the arguments about relationship to expression were ignoratio elenchi. I agree. I assert, however, that my statement that a question is a speech act ("operation") involving a concept or proposition must then be what was intended. As a speech act, it differs from an assertion, for example, but not from a "proposition", which (we say) is a meaning, something acted on by a speech act. My main concern was to get the SBVR subtypes of 'meaning' to be related by some principle. And I can't find the principle that puts questions in the same taxonomy with concept and proposition. > * I am fine with your "3 kinds of questions". But this seems to > vitiate your suggested, new definition of question ("a meaning that > explicitly refers to the extension of a concept"), since that definition > only seems relevant to "What" questions (those which ask "for the > extension of a concept"). Am I missing something here? I only proposed that because it matches Don's 'question formulation' in clause 9, which, according to Don, treats Whether questions as projections whose result space is the powerset of {true, false}. Put simply, I don't know what to do with questions, but the current text of SBVR once again demonstrates Pope's aphorism. Since you do seem to know what the relationship of questions to meanings is, perhaps you could suggest a solution. (And if "no change" is the solution, it resolves the issue.) Thanks, -Ed -- National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." Subject: RE: issue 10621 -- SBVR FTF issue Date: Wed, 24 Jan 2007 17:29:34 -0800 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: issue 10621 -- SBVR FTF issue Thread-Index: AcdABygWwT5QSRYaSM+aFIO6IIQatQADOl9wAALTrRA= From: "Baisley, Donald E" To: "Andy Carver" , Cc: X-OriginalArrivalTime: 25 Jan 2007 01:29:55.0989 (UTC) FILETIME=[51DC9450:01C74020] X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id l0P1MHia024372 Hi Andy and Ed, 'question' as SBVR has it, is supposed to be independent of any speech act. John asked whether snow is cold. John answered whether snow is cold. John decided whether snow is cold. These statements all nominalize the same question, but they are about different acts. The same question can be stated in different languages, just as a proposition can be stated in different languages. The 'question' we are talking about in SBVR is the one meaning. Regards, Don -----Original Message----- From: Andy Carver [mailto:Andy.Carver@neumont.edu] Sent: Wednesday, January 24, 2007 4:25 PM To: edbark@nist.gov Cc: sbvr-ftf@omg.org Subject: RE: issue 10621 -- SBVR FTF issue Ed, You wrote, "My main concern was to get the SBVR subtypes of 'meaning' to be related by some principle. And I can't find the principle that puts questions in the same taxonomy with concept and proposition." I must admit I hadn't noticed that very interesting and important problem. And now that I've pointed out that "question" is a kind of speech act, it makes the problem appear even more acute; because one definitely cannot say this same thing about "proposition" or "concept". So are we mixing apples and oranges, and making a taxonomy error? You've got me wondering too. Speech acts often relate to propositions by having a propositional component, as well as a performative component: for example, the sort of speech act called "assertion" would have a meaning-component, viz. the proposition being asserted as true (also called the "locutionary" component), and a performative component, the act/attitude of asserting it. Likewise would a "true/false" question; but here the performative component would obviously be a different one. "What" questions, as you called them, would NOT have a propositional component, since they're not asking for a truth value but for an extension. But perhaps we could say they still have a MEANING-component, viz. the concept whose extension they're requesting. But I'm not sure whether standard speech-act theory allows for such thin spreading of the idea of the "locutionary meaning"; I'd have to research that. In any case, the interesting thing right now is that speech acts are not the meanings, but rather, they contain such "locutionary" meanings but also "performative" and "perlocutionary" meanings. So at this point, it seems to me you've identified a real quandary in our taxonomy. Wish I could shed more light on it, but right now that's the extent of my wisdom. Cheers, Andy -----Original Message----- From: Ed Barkmeyer [mailto:edbark@nist.gov] Sent: Wednesday, January 24, 2007 3:30 PM To: Andy Carver Cc: sbvr-ftf@omg.org Subject: Re: issue 10621 -- SBVR FTF issue Andy Carver wrote: > * Your issue is "assuming that 'interrogatory' means > 'interrogatory sentence'". And yet I would think this assumption is > mistaken. It is not an interrogatory sentence that is in view, but an > interrogatory speech act, in the technical sense intended in "speech act > theory". And the whole problem statement, it seems to me, assumes the > former definition of "interrogatory"; because once you accept the latter > instead, I don't see that context affects the meaning. Thank you. Neither my ODE nor my Webster defines to "interrogatory" as a term in speech act theory. The essence of both definitions is that the noun 'interrogatory' means a written question that requires a formal response. According to SBVR, therefore, that is the meaning. We apparently need at least a minor editorial change to associate it with speech act theory. At the same time, my assumption that it meant an interrogative sentence was unwarranted. And the arguments about relationship to expression were ignoratio elenchi. I agree. I assert, however, that my statement that a question is a speech act ("operation") involving a concept or proposition must then be what was intended. As a speech act, it differs from an assertion, for example, but not from a "proposition", which (we say) is a meaning, something acted on by a speech act. My main concern was to get the SBVR subtypes of 'meaning' to be related by some principle. And I can't find the principle that puts questions in the same taxonomy with concept and proposition. > * I am fine with your "3 kinds of questions". But this seems to > vitiate your suggested, new definition of question ("a meaning that > explicitly refers to the extension of a concept"), since that definition > only seems relevant to "What" questions (those which ask "for the > extension of a concept"). Am I missing something here? I only proposed that because it matches Don's 'question formulation' in clause 9, which, according to Don, treats Whether questions as projections whose result space is the powerset of {true, false}. Put simply, I don't know what to do with questions, but the current text of SBVR once again demonstrates Pope's aphorism. Since you do seem to know what the relationship of questions to meanings is, perhaps you could suggest a solution. (And if "no change" is the solution, it resolves the issue.) Thanks, -Ed -- National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." Date: Thu, 25 Jan 2007 11:09:03 -0500 From: Ed Barkmeyer Reply-To: edbark@nist.gov Organization: NIST User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax) X-Accept-Language: en, fr, de, pdf, it, nl, sv, es, ru To: "Baisley, Donald E" CC: sbvr-ftf@omg.org Subject: Re: issue 10621 -- SBVR FTF issue X-MailScanner-Information: Please contact postmaster@mel.nist.gov for more information X-MailScanner: Found to be clean X-MailScanner-SpamCheck: X-MailScanner-From: edbark@nist.gov X-Spam-Status: No Baisley, Donald E wrote: 'question' as SBVR has it, is supposed to be independent of any speech act. John asked whether snow is cold. John answered whether snow is cold. John decided whether snow is cold. These statements all nominalize the same question, but they are about different acts. The same question can be stated in different languages, just as a proposition can be stated in different languages. The 'question' we are talking about in SBVR is the one meaning. But that meaning is a speech act applied to the *proposition* 'snow is cold'. I don't have a background in speech act theory, so I don't know how many of the above may be "interrogatories", but they all operate on that proposition, as does: John said that snow is cold. Everyone believes that snow is cold. Snow is cold. All of these are also speech acts involving that proposition. (The last may be interpreted: This email asserts that snow is cold.) Further, in John asked p, and John decided p, the 'whether' is an accident of the English language phrasing of the speech act, just as 'John asserted p' can be phrased 'John said that p', in which 'that' is an accident of the natural language. And I believe the same argument can be made for 'John answered whether p', which is interpreted: John said that the truth-value of p is v, where John knew/believed v to be the truth value of p. So if 'question' were independent of all speech acts, it would be indistinguishable from 'proposition'. And that was my point in saying that, as Andy put it, we are mixing apples and oranges in our taxonomy of meaning. -Ed -- Edward J. Barkmeyer Email: edbark@nist.gov National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." Date: Thu, 25 Jan 2007 19:00:52 -0500 From: Ed Barkmeyer Reply-To: edbark@nist.gov Organization: NIST User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax) X-Accept-Language: en, fr, de, pdf, it, nl, sv, es, ru To: "Baisley, Donald E" CC: sbvr-ftf@omg.org Subject: Re: [SBVR-FTF] issue 10621 -- Questions X-MailScanner-Information: Please contact postmaster@mel.nist.gov for more information X-MailScanner: Found to be clean X-MailScanner-SpamCheck: X-MailScanner-From: edbark@nist.gov X-Spam-Status: No Baisley, Donald E wrote: So if 'question' were independent of all speech acts, it would be indistinguishable from 'proposition'. "whether snow is cold" means "what truth-value is had by the proposition that snow is cold". The meaning of "whether snow is cold" is not the same as "snow is cold". But in this sense, asserting "that snow is cold" is not the same as "snow is cold", either. The meaning of the assertion "that snow is cold" is "the truth value of the proposition 'snow is cold' is 'true'." This is the problem. A question is different from a proposition in that it is a speech act that operates on a proposition. An assertion is different from a proposition in that it is a speech act that operates on a proposition. They both differ in meaning from the proposition itself by the fact of the speech act. And they differ from each other in the nature of the speech act. Questions, assertions, denials, etc. are all speech acts that operate on propositions. Those speech acts add meaning. The point of 10621 is that the taxonomy of 'meaning' might appropriately be 'concept', 'proposition', 'speech act' (I'm not sure of that), but 'question' is just one subtype of 'speech act', and ALL speech acts are different from propositions and concepts. Propositions involve concepts, and speech acts involve propositions. I will use another example question. ... ... A question is a meaning, similar to a proposition, formed in the mind. Like a proposition, it can have multiple representations, but it is not a proposition. And we should not confuse a question with the acts of asking it. Yes, a question is a meaning. I don't know enough about speech act theory to say, but I think that the term speech act refers to the meaning that is added to a proposition when the proposition is used in communication. It is not the physical act of asking a question that makes it a question. It is the conceptual act of evaluating a proposition, as distinct from the act of asserting the proposition or denying it or mentioning it. Put another way, when I was a debater in high school and college, the stage would be set with words like: Today we will debate the proposition "Resolved, that the United Nations has improved the African economies." Debating the proposition means discussing whether the proposition holds. The side called the Affirmation asserts the proposition; the side called the Negation denies the proposition. Debating or discussing a proposition is similar to "asking the question". It is not clear that anyone "asks the question", but they all act on the proposition. The point is that there is nothing special about 'question' here -- there are several different acts on the proposition, and posing the proposition as a 'question' is just one of them. Issue 10621 is in part whether 'question' is the parallel to 'proposition' and 'concept'. And I think it is, but it is only an example of meanings that are parallel to 'proposition'. An assertion is a meaning, a denial is a meaning, an imperative is a meaning. They are all different from proposition and concept and from question. Where do they appear in the SBVR taxonomy? How does SBVR model them? As mentions? What is special about question? I also think SBVR has rolled up two somehow different ideas in the term 'question': the evaluation of a proposition, and the enumeration of the extension of a concept. They act on different things. I don't know where this takes us. Issue 10621 says in effect that we have one term for two concepts, and one of those concepts -- the speech act -- is a proper subtype of a class of meanings we might want to include in SBVR. I wish I knew what theory should be applied to resolve this concern. I don't. -Ed -- Edward J. Barkmeyer Email: edbark@nist.gov National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." Subject: RE: [SBVR-FTF] issue 10621 -- Questions Date: Thu, 25 Jan 2007 17:04:19 -0700 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: [SBVR-FTF] issue 10621 -- Questions Thread-Index: AcdAwUVLS7KRMeBUTcGjAfspZPqfRwAG/M7A From: "Andy Carver" To: "Baisley, Donald E" , Cc: X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id l0PNuZtG013014 Don, Some of your comments on your new example show that you're slipping on the technical sense of "speech act". #1 involves exactly one "speech act" in the technical sense, and so does #2. "Speech act" does NOT mean "the acts of asking it". Andy -----Original Message----- From: Baisley, Donald E [mailto:Donald.Baisley@unisys.com] Sent: Thursday, January 25, 2007 1:40 PM To: edbark@nist.gov Cc: sbvr-ftf@omg.org Subject: RE: [SBVR-FTF] issue 10621 -- Questions Ed wrote: > So if 'question' were independent of all speech acts, > it would be indistinguishable from 'proposition'. "whether snow is cold" means "what truth-value is had by the proposition that snow is cold". The meaning of "whether snow is cold" is not the same as "snow is cold". I will use another example question. 1. John asked what time Sally came, and Ralph asked the same question. 2. The question of what time Sally came popped into Fred's mind, but was never asked. 3. The question of what time Sally came was answered incorrectly by Joe. 4. Mark determined the answer to what time Sally came by driving her in his car. 5. The question of what time Sally came was expressed in Spanish by a Jose. While #1 talks of speech acts, two askings of the same question, #2 involves no speech act. #5 shows that a question is not tied to the words used to expressed it. A question is a meaning, similar to a proposition, formed in the mind. Like a proposition, it can have multiple representations, but it is not a proposition. And we should not confuse a question with the acts of asking it. Enjoy, Don Subject: RE: [SBVR-FTF] issue 10621 -- Questions Date: Thu, 25 Jan 2007 17:14:04 -0700 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: [SBVR-FTF] issue 10621 -- Questions Thread-Index: AcdA3TsUFUat2SFuRI2stecs/yaU9gAANd1Q From: "Andy Carver" To: , "Baisley, Donald E" Cc: X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id l0Q06OLc018267 Ed and Don, I agree with basically all that Ed said below. To make more concrete both the problem and direction it might take us, let me suggest the following possible taxonomy: Meaning: - Locutionary meaning: -- Proposition -- Concept - Performative meaning: -- Question: --- "What" question --- "Whether" question - Perlocutionary meaning This might well be improvable, but at least it's making some needed but currently missing distinctions. Regards, Andy -----Original Message----- From: Ed Barkmeyer [mailto:edbark@nist.gov] Sent: Thursday, January 25, 2007 5:01 PM To: Baisley, Donald E Cc: sbvr-ftf@omg.org Subject: Re: [SBVR-FTF] issue 10621 -- Questions Baisley, Donald E wrote: >>So if 'question' were independent of all speech acts, >>it would be indistinguishable from 'proposition'. > > "whether snow is cold" means "what truth-value is had by the proposition > that snow is cold". The meaning of "whether snow is cold" is not the > same as "snow is cold". But in this sense, asserting "that snow is cold" is not the same as "snow is cold", either. The meaning of the assertion "that snow is cold" is "the truth value of the proposition 'snow is cold' is 'true'." This is the problem. A question is different from a proposition in that it is a speech act that operates on a proposition. An assertion is different from a proposition in that it is a speech act that operates on a proposition. They both differ in meaning from the proposition itself by the fact of the speech act. And they differ from each other in the nature of the speech act. Questions, assertions, denials, etc. are all speech acts that operate on propositions. Those speech acts add meaning. The point of 10621 is that the taxonomy of 'meaning' might appropriately be 'concept', 'proposition', 'speech act' (I'm not sure of that), but 'question' is just one subtype of 'speech act', and ALL speech acts are different from propositions and concepts. Propositions involve concepts, and speech acts involve propositions. > I will use another example question. ... > ... A question is a meaning, similar to a > proposition, formed in the mind. Like a proposition, it can have > multiple representations, but it is not a proposition. And we should > not confuse a question with the acts of asking it. Yes, a question is a meaning. I don't know enough about speech act theory to say, but I think that the term speech act refers to the meaning that is added to a proposition when the proposition is used in communication. It is not the physical act of asking a question that makes it a question. It is the conceptual act of evaluating a proposition, as distinct from the act of asserting the proposition or denying it or mentioning it. Put another way, when I was a debater in high school and college, the stage would be set with words like: Today we will debate the proposition "Resolved, that the United Nations has improved the African economies." Debating the proposition means discussing whether the proposition holds. The side called the Affirmation asserts the proposition; the side called the Negation denies the proposition. Debating or discussing a proposition is similar to "asking the question". It is not clear that anyone "asks the question", but they all act on the proposition. The point is that there is nothing special about 'question' here -- there are several different acts on the proposition, and posing the proposition as a 'question' is just one of them. Issue 10621 is in part whether 'question' is the parallel to 'proposition' and 'concept'. And I think it is, but it is only an example of meanings that are parallel to 'proposition'. An assertion is a meaning, a denial is a meaning, an imperative is a meaning. They are all different from proposition and concept and from question. Where do they appear in the SBVR taxonomy? How does SBVR model them? As mentions? What is special about question? I also think SBVR has rolled up two somehow different ideas in the term 'question': the evaluation of a proposition, and the enumeration of the extension of a concept. They act on different things. I don't know where this takes us. Issue 10621 says in effect that we have one term for two concepts, and one of those concepts -- the speech act -- is a proper subtype of a class of meanings we might want to include in SBVR. I wish I knew what theory should be applied to resolve this concern. I don't. -Ed -- National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." Subject: RE: [SBVR-FTF] issue 10621 -- Questions Date: Thu, 25 Jan 2007 17:50:01 -0800 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: [SBVR-FTF] issue 10621 -- Questions Thread-Index: AcdA3TsUFUat2SFuRI2stecs/yaU9gAANd1QAAK2hpA= From: "Baisley, Donald E" To: "Andy Carver" , Cc: X-OriginalArrivalTime: 26 Jan 2007 01:50:03.0024 (UTC) FILETIME=[4BB94100:01C740EC] X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id l0Q1gQen030461 Andy's taxonomy shows 'proposition', 'concept' and 'question' all specializing the more general concept 'meaning'. It also shows several other concepts that are not part of SBVR. If we agree with Andy, then we agree that SBVR's taxonomy of 'meaning', 'proposition', 'concept' and 'question' is correct as it stands. SBVR makes no claim that its categories of 'meaning' are complete (which would be an absurd claim), nor does it claim to address all layers that might appear in a complete taxonomy conceived by linguists, philosophers and logicians. SBVR does not need to expand to become complete in that area. I don't understand Ed's points about an "act" being a meaning, but I don't think SBVR needs to go there. If there is some change to the wording of the definition of 'question' that would make it more clear, then let's do it. Don -----Original Message----- From: Andy Carver [mailto:Andy.Carver@neumont.edu] Sent: Thursday, January 25, 2007 4:14 PM To: edbark@nist.gov; Baisley, Donald E Cc: sbvr-ftf@omg.org Subject: RE: [SBVR-FTF] issue 10621 -- Questions Ed and Don, I agree with basically all that Ed said below. To make more concrete both the problem and direction it might take us, let me suggest the following possible taxonomy: Meaning: - Locutionary meaning: -- Proposition -- Concept - Performative meaning: -- Question: --- "What" question --- "Whether" question - Perlocutionary meaning This might well be improvable, but at least it's making some needed but currently missing distinctions. Regards, Andy -----Original Message----- From: Ed Barkmeyer [mailto:edbark@nist.gov] Sent: Thursday, January 25, 2007 5:01 PM To: Baisley, Donald E Cc: sbvr-ftf@omg.org Subject: Re: [SBVR-FTF] issue 10621 -- Questions Baisley, Donald E wrote: >>So if 'question' were independent of all speech acts, >>it would be indistinguishable from 'proposition'. > > "whether snow is cold" means "what truth-value is had by the proposition > that snow is cold". The meaning of "whether snow is cold" is not the > same as "snow is cold". But in this sense, asserting "that snow is cold" is not the same as "snow is cold", either. The meaning of the assertion "that snow is cold" is "the truth value of the proposition 'snow is cold' is 'true'." This is the problem. A question is different from a proposition in that it is a speech act that operates on a proposition. An assertion is different from a proposition in that it is a speech act that operates on a proposition. They both differ in meaning from the proposition itself by the fact of the speech act. And they differ from each other in the nature of the speech act. Questions, assertions, denials, etc. are all speech acts that operate on propositions. Those speech acts add meaning. The point of 10621 is that the taxonomy of 'meaning' might appropriately be 'concept', 'proposition', 'speech act' (I'm not sure of that), but 'question' is just one subtype of 'speech act', and ALL speech acts are different from propositions and concepts. Propositions involve concepts, and speech acts involve propositions. > I will use another example question. ... > ... A question is a meaning, similar to a > proposition, formed in the mind. Like a proposition, it can have > multiple representations, but it is not a proposition. And we should > not confuse a question with the acts of asking it. Yes, a question is a meaning. I don't know enough about speech act theory to say, but I think that the term speech act refers to the meaning that is added to a proposition when the proposition is used in communication. It is not the physical act of asking a question that makes it a question. It is the conceptual act of evaluating a proposition, as distinct from the act of asserting the proposition or denying it or mentioning it. Put another way, when I was a debater in high school and college, the stage would be set with words like: Today we will debate the proposition "Resolved, that the United Nations has improved the African economies." Debating the proposition means discussing whether the proposition holds. The side called the Affirmation asserts the proposition; the side called the Negation denies the proposition. Debating or discussing a proposition is similar to "asking the question". It is not clear that anyone "asks the question", but they all act on the proposition. The point is that there is nothing special about 'question' here -- there are several different acts on the proposition, and posing the proposition as a 'question' is just one of them. Issue 10621 is in part whether 'question' is the parallel to 'proposition' and 'concept'. And I think it is, but it is only an example of meanings that are parallel to 'proposition'. An assertion is a meaning, a denial is a meaning, an imperative is a meaning. They are all different from proposition and concept and from question. Where do they appear in the SBVR taxonomy? How does SBVR model them? As mentions? What is special about question? I also think SBVR has rolled up two somehow different ideas in the term 'question': the evaluation of a proposition, and the enumeration of the extension of a concept. They act on different things. I don't know where this takes us. Issue 10621 says in effect that we have one term for two concepts, and one of those concepts -- the speech act -- is a proper subtype of a class of meanings we might want to include in SBVR. I wish I knew what theory should be applied to resolve this concern. I don't. -Ed -- National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." Subject: RE: [SBVR-FTF] issue 10621 -- Questions Date: Thu, 25 Jan 2007 19:28:38 -0700 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: [SBVR-FTF] issue 10621 -- Questions Thread-Index: AcdA7FO5cjLWaVpmT3iDZHrOKL2BmQABVxH3 From: "Andy Carver" To: "Baisley, Donald E" , Cc: Don, Yes -- but not so fast: I just threw that out as a possibility. Even as I was doing it, I was consciously doubtful that you could really say that a "question" was a kind of performative meaning. It seems to me that you can definitely use the term "'whether' question" to refer to a definite, particular sort of speech act. I'm much more sure of that, than I am that we can use the same term to mean a kind of performative meaning. I'm hoping for Ed's feedback on this latter possibility, but I consider it guilty till proven innocent. Andy -------------------------------------------------------------------------------- From: Baisley, Donald E [mailto:Donald.Baisley@unisys.com] Sent: Thu 1/25/2007 6:50 PM To: Andy Carver; edbark@nist.gov Cc: sbvr-ftf@omg.org Subject: RE: [SBVR-FTF] issue 10621 -- Questions Andy's taxonomy shows 'proposition', 'concept' and 'question' all specializing the more general concept 'meaning'. It also shows several other concepts that are not part of SBVR. If we agree with Andy, then we agree that SBVR's taxonomy of 'meaning', 'proposition', 'concept' and 'question' is correct as it stands. SBVR makes no claim that its categories of 'meaning' are complete (which would be an absurd claim), nor does it claim to address all layers that might appear in a complete taxonomy conceived by linguists, philosophers and logicians. SBVR does not need to expand to become complete in that area. I don't understand Ed's points about an "act" being a meaning, but I don't think SBVR needs to go there. If there is some change to the wording of the definition of 'question' that would make it more clear, then let's do it. Don -----Original Message----- From: Andy Carver [mailto:Andy.Carver@neumont.edu] Sent: Thursday, January 25, 2007 4:14 PM To: edbark@nist.gov; Baisley, Donald E Cc: sbvr-ftf@omg.org Subject: RE: [SBVR-FTF] issue 10621 -- Questions Ed and Don, I agree with basically all that Ed said below. To make more concrete both the problem and direction it might take us, let me suggest the following possible taxonomy: Meaning: - Locutionary meaning: -- Proposition -- Concept - Performative meaning: -- Question: --- "What" question --- "Whether" question - Perlocutionary meaning This might well be improvable, but at least it's making some needed but currently missing distinctions. Regards, Andy -----Original Message----- From: Ed Barkmeyer [mailto:edbark@nist.gov] Sent: Thursday, January 25, 2007 5:01 PM To: Baisley, Donald E Cc: sbvr-ftf@omg.org Subject: Re: [SBVR-FTF] issue 10621 -- Questions Baisley, Donald E wrote: >>So if 'question' were independent of all speech acts, >>it would be indistinguishable from 'proposition'. > > "whether snow is cold" means "what truth-value is had by the proposition > that snow is cold". The meaning of "whether snow is cold" is not the > same as "snow is cold". But in this sense, asserting "that snow is cold" is not the same as "snow is cold", either. The meaning of the assertion "that snow is cold" is "the truth value of the proposition 'snow is cold' is 'true'." This is the problem. A question is different from a proposition in that it is a speech act that operates on a proposition. An assertion is different from a proposition in that it is a speech act that operates on a proposition. They both differ in meaning from the proposition itself by the fact of the speech act. And they differ from each other in the nature of the speech act. Questions, assertions, denials, etc. are all speech acts that operate on propositions. Those speech acts add meaning. The point of 10621 is that the taxonomy of 'meaning' might appropriately be 'concept', 'proposition', 'speech act' (I'm not sure of that), but 'question' is just one subtype of 'speech act', and ALL speech acts are different from propositions and concepts. Propositions involve concepts, and speech acts involve propositions. > I will use another example question. ... > ... A question is a meaning, similar to a > proposition, formed in the mind. Like a proposition, it can have > multiple representations, but it is not a proposition. And we should > not confuse a question with the acts of asking it. Yes, a question is a meaning. I don't know enough about speech act theory to say, but I think that the term speech act refers to the meaning that is added to a proposition when the proposition is used in communication. It is not the physical act of asking a question that makes it a question. It is the conceptual act of evaluating a proposition, as distinct from the act of asserting the proposition or denying it or mentioning it. Put another way, when I was a debater in high school and college, the stage would be set with words like: Today we will debate the proposition "Resolved, that the United Nations has improved the African economies." Debating the proposition means discussing whether the proposition holds. The side called the Affirmation asserts the proposition; the side called the Negation denies the proposition. Debating or discussing a proposition is similar to "asking the question". It is not clear that anyone "asks the question", but they all act on the proposition. The point is that there is nothing special about 'question' here -- there are several different acts on the proposition, and posing the proposition as a 'question' is just one of them. Issue 10621 is in part whether 'question' is the parallel to 'proposition' and 'concept'. And I think it is, but it is only an example of meanings that are parallel to 'proposition'. An assertion is a meaning, a denial is a meaning, an imperative is a meaning. They are all different from proposition and concept and from question. Where do they appear in the SBVR taxonomy? How does SBVR model them? As mentions? What is special about question? I also think SBVR has rolled up two somehow different ideas in the term 'question': the evaluation of a proposition, and the enumeration of the extension of a concept. They act on different things. I don't know where this takes us. Issue 10621 says in effect that we have one term for two concepts, and one of those concepts -- the speech act -- is a proper subtype of a class of meanings we might want to include in SBVR. I wish I knew what theory should be applied to resolve this concern. I don't. -Ed -- National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." Subject: RE: [SBVR-FTF] issue 10621 -- Questions Date: Thu, 25 Jan 2007 19:47:12 -0700 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: [SBVR-FTF] issue 10621 -- Questions Thread-Index: AcdA7FO5cjLWaVpmT3iDZHrOKL2BmQABVxH3AABq7hA= From: "Andy Carver" To: "Baisley, Donald E" , Cc: Don and Ed, Here's another possibility, for whose potential I carry much more optimism than I ever did for the last one's: Speech Act has Locutionary Meaning; Speech Act has Illocutionary (a.k.a. "Performative") Meaning; Speech Act has Perlocutionary Meaning; Locutionary Meaning: - Concept - Proposition Speech Act: - Question -- "Whether" Question -- "What" Question As before, feedback desired. Andy -------------------------------------------------------------------------------- From: Andy Carver Sent: Thu 1/25/2007 7:28 PM To: Baisley, Donald E; edbark@nist.gov Cc: sbvr-ftf@omg.org Subject: RE: [SBVR-FTF] issue 10621 -- Questions Don, Yes -- but not so fast: I just threw that out as a possibility. Even as I was doing it, I was consciously doubtful that you could really say that a "question" was a kind of performative meaning. It seems to me that you can definitely use the term "'whether' question" to refer to a definite, particular sort of speech act. I'm much more sure of that, than I am that we can use the same term to mean a kind of performative meaning. I'm hoping for Ed's feedback on this latter possibility, but I consider it guilty till proven innocent. Andy -------------------------------------------------------------------------------- From: Baisley, Donald E [mailto:Donald.Baisley@unisys.com] Sent: Thu 1/25/2007 6:50 PM To: Andy Carver; edbark@nist.gov Cc: sbvr-ftf@omg.org Subject: RE: [SBVR-FTF] issue 10621 -- Questions Andy's taxonomy shows 'proposition', 'concept' and 'question' all specializing the more general concept 'meaning'. It also shows several other concepts that are not part of SBVR. If we agree with Andy, then we agree that SBVR's taxonomy of 'meaning', 'proposition', 'concept' and 'question' is correct as it stands. SBVR makes no claim that its categories of 'meaning' are complete (which would be an absurd claim), nor does it claim to address all layers that might appear in a complete taxonomy conceived by linguists, philosophers and logicians. SBVR does not need to expand to become complete in that area. I don't understand Ed's points about an "act" being a meaning, but I don't think SBVR needs to go there. If there is some change to the wording of the definition of 'question' that would make it more clear, then let's do it. Don -----Original Message----- From: Andy Carver [mailto:Andy.Carver@neumont.edu] Sent: Thursday, January 25, 2007 4:14 PM To: edbark@nist.gov; Baisley, Donald E Cc: sbvr-ftf@omg.org Subject: RE: [SBVR-FTF] issue 10621 -- Questions Ed and Don, I agree with basically all that Ed said below. To make more concrete both the problem and direction it might take us, let me suggest the following possible taxonomy: Meaning: - Locutionary meaning: -- Proposition -- Concept - Performative meaning: -- Question: --- "What" question --- "Whether" question - Perlocutionary meaning This might well be improvable, but at least it's making some needed but currently missing distinctions. Regards, Andy -----Original Message----- From: Ed Barkmeyer [mailto:edbark@nist.gov] Sent: Thursday, January 25, 2007 5:01 PM To: Baisley, Donald E Cc: sbvr-ftf@omg.org Subject: Re: [SBVR-FTF] issue 10621 -- Questions Baisley, Donald E wrote: >>So if 'question' were independent of all speech acts, >>it would be indistinguishable from 'proposition'. > > "whether snow is cold" means "what truth-value is had by the proposition > that snow is cold". The meaning of "whether snow is cold" is not the > same as "snow is cold". But in this sense, asserting "that snow is cold" is not the same as "snow is cold", either. The meaning of the assertion "that snow is cold" is "the truth value of the proposition 'snow is cold' is 'true'." This is the problem. A question is different from a proposition in that it is a speech act that operates on a proposition. An assertion is different from a proposition in that it is a speech act that operates on a proposition. They both differ in meaning from the proposition itself by the fact of the speech act. And they differ from each other in the nature of the speech act. Questions, assertions, denials, etc. are all speech acts that operate on propositions. Those speech acts add meaning. The point of 10621 is that the taxonomy of 'meaning' might appropriately be 'concept', 'proposition', 'speech act' (I'm not sure of that), but 'question' is just one subtype of 'speech act', and ALL speech acts are different from propositions and concepts. Propositions involve concepts, and speech acts involve propositions. > I will use another example question. ... > ... A question is a meaning, similar to a > proposition, formed in the mind. Like a proposition, it can have > multiple representations, but it is not a proposition. And we should > not confuse a question with the acts of asking it. Yes, a question is a meaning. I don't know enough about speech act theory to say, but I think that the term speech act refers to the meaning that is added to a proposition when the proposition is used in communication. It is not the physical act of asking a question that makes it a question. It is the conceptual act of evaluating a proposition, as distinct from the act of asserting the proposition or denying it or mentioning it. Put another way, when I was a debater in high school and college, the stage would be set with words like: Today we will debate the proposition "Resolved, that the United Nations has improved the African economies." Debating the proposition means discussing whether the proposition holds. The side called the Affirmation asserts the proposition; the side called the Negation denies the proposition. Debating or discussing a proposition is similar to "asking the question". It is not clear that anyone "asks the question", but they all act on the proposition. The point is that there is nothing special about 'question' here -- there are several different acts on the proposition, and posing the proposition as a 'question' is just one of them. Issue 10621 is in part whether 'question' is the parallel to 'proposition' and 'concept'. And I think it is, but it is only an example of meanings that are parallel to 'proposition'. An assertion is a meaning, a denial is a meaning, an imperative is a meaning. They are all different from proposition and concept and from question. Where do they appear in the SBVR taxonomy? How does SBVR model them? As mentions? What is special about question? I also think SBVR has rolled up two somehow different ideas in the term 'question': the evaluation of a proposition, and the enumeration of the extension of a concept. They act on different things. I don't know where this takes us. Issue 10621 says in effect that we have one term for two concepts, and one of those concepts -- the speech act -- is a proper subtype of a class of meanings we might want to include in SBVR. I wish I knew what theory should be applied to resolve this concern. I don't. -Ed -- National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=btinternet.com; h=Received:X-YMail-OSG:Reply-To:From:To:References:Subject:Date:Organization:Message-ID:MIME-Version:Content-Type:X-Mailer:X-MimeOLE:Thread-Index:In-Reply-To; b=ecDPe3uf/f1LHUd5LfQy3g5jTAM7LdtuhwWUki8YkvssAwdKaV04N9wdmz6iIMO5wXDcsiBdAcLBOUcbWbEXwjhialFg/jiqpiC+n3/Du9IcNg79Mps+1GXIYImin+NJSYdEvfkK6+l27QUyxWDHk36z3GnB4oTB04KKPlbi+iU= ; X-YMail-OSG: Gw5bGrkVM1mB738ejAzyfXl2why0Gh0LfZ2QIL6w4gOLu5Bk7.qthvTh818ZnTjmQH_9KBN_bm0DGGlfWiN3GdaetCFakbFi6rB1.Htjew-- Reply-To: From: "Donald Chapin" To: Subject: RE: [SBVR-FTF] issue 10621 -- Questions Date: Fri, 26 Jan 2007 15:11:12 -0000 Organization: Business Semantics Ltd X-Mailer: Microsoft Office Outlook 11 Thread-Index: AcdA3RPzC0LcrSL+TnmOXey476vp5gAe2kJQ Ed, Don & Andy, The "performative (illocutionary force)| + proposition" construct has been a fundamental part of the SBVR architecture from the beginning (BRT decision at Sept. 2003 OMG Meeting in Boston). Attached is the latest working document (dec. 2003) that took a total SBVR architectural view, as well as a slide that summarizes "proposition + performative" (taken from the Boston meeting presentations). Dealing with proposition separately from illocutionary force and dealing with "proposition + illocutionary force" was a key reason for the UNCEFACT excitement about SBVR. Illocutionary force (speech acts) is a key element in the ISO TC 37 SC 4 language parsing standards which will need to be connectable with SBVR. While we clearly have not fully set forth the "performative + proposition" architecture in the current SBVR document (and should not try to do so before finalization), nothing that is in SBVR should contradict it. We should neither conflate nor confuse the constructs for 'proposition', 'performative (illocutionary force, speech act)' or 'proposition + performative'. Donald > -----Original Message----- > From: Ed Barkmeyer [mailto:edbark@nist.gov] > Sent: 26 January 2007 00:01 > To: Baisley, Donald E > Cc: sbvr-ftf@omg.org > Subject: Re: [SBVR-FTF] issue 10621 -- Questions > > Baisley, Donald E wrote: > > >>So if 'question' were independent of all speech acts, > >>it would be indistinguishable from 'proposition'. > > > > "whether snow is cold" means "what truth-value is had by the proposition > > that snow is cold". The meaning of "whether snow is cold" is not the > > same as "snow is cold". > > But in this sense, asserting "that snow is cold" is not the same as "snow > is > cold", either. The meaning of the assertion "that snow is cold" is "the > truth > value of the proposition 'snow is cold' is 'true'." > > This is the problem. A question is different from a proposition in that > it is > a speech act that operates on a proposition. An assertion is different > from a > proposition in that it is a speech act that operates on a proposition. > They > both differ in meaning from the proposition itself by the fact of the > speech > act. And they differ from each other in the nature of the speech act. > > Questions, assertions, denials, etc. are all speech acts that operate on > propositions. Those speech acts add meaning. The point of 10621 is that > the > taxonomy of 'meaning' might appropriately be 'concept', 'proposition', > 'speech > act' (I'm not sure of that), but 'question' is just one subtype of 'speech > act', and ALL speech acts are different from propositions and concepts. > Propositions involve concepts, and speech acts involve propositions. > > > I will use another example question. ... > > > ... A question is a meaning, similar to a > > proposition, formed in the mind. Like a proposition, it can have > > multiple representations, but it is not a proposition. And we should > > not confuse a question with the acts of asking it. > > Yes, a question is a meaning. I don't know enough about speech act theory > to > say, but I think that the term speech act refers to the meaning that is > added > to a proposition when the proposition is used in communication. It is not > the > physical act of asking a question that makes it a question. It is the > conceptual act of evaluating a proposition, as distinct from the act of > asserting the proposition or denying it or mentioning it. > > Put another way, when I was a debater in high school and college, the > stage > would be set with words like: Today we will debate the proposition > "Resolved, > that the United Nations has improved the African economies." Debating the > proposition means discussing whether the proposition holds. The side > called > the Affirmation asserts the proposition; the side called the Negation > denies > the proposition. Debating or discussing a proposition is similar to > "asking > the question". It is not clear that anyone "asks the question", but they > all > act on the proposition. The point is that there is nothing special about > 'question' here -- there are several different acts on the proposition, > and > posing the proposition as a 'question' is just one of them. > > Issue 10621 is in part whether 'question' is the parallel to 'proposition' > and > 'concept'. And I think it is, but it is only an example of meanings that > are > parallel to 'proposition'. An assertion is a meaning, a denial is a > meaning, > an imperative is a meaning. They are all different from proposition and > concept and from question. Where do they appear in the SBVR taxonomy? > How > does SBVR model them? As mentions? What is special about question? > > I also think SBVR has rolled up two somehow different ideas in the term > 'question': the evaluation of a proposition, and the enumeration of the > extension of a concept. They act on different things. I don't know where > this takes us. > > Issue 10621 says in effect that we have one term for two concepts, and one > of > those concepts -- the speech act -- is a proper subtype of a class of > meanings > we might want to include in SBVR. > > I wish I knew what theory should be applied to resolve this concern. I > don't. > > -Ed > > -- > Edward J. Barkmeyer Email: edbark@nist.gov > National Institute of Standards & Technology > Manufacturing Systems Integration Division > 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 > Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 > > "The opinions expressed above do not reflect consensus of NIST, > and have not been reviewed by any Government authority." Illocutionary_Force_Chart_--_Working_Document.doc Propositional Content + Performative.ppt Subject: RE: [SBVR-FTF] issue 10621 -- Questions Date: Thu, 25 Jan 2007 12:39:32 -0800 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: [SBVR-FTF] issue 10621 -- Questions Thread-Index: AcdAmzJRO0oNqvdxT3y7Ugc7fCatlgAGrlpw From: "Baisley, Donald E" To: Cc: X-OriginalArrivalTime: 25 Jan 2007 20:39:44.0248 (UTC) FILETIME=[F2117F80:01C740C0] X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id l0PKW22g010507 Ed wrote: > So if 'question' were independent of all speech acts, > it would be indistinguishable from 'proposition'. "whether snow is cold" means "what truth-value is had by the proposition that snow is cold". The meaning of "whether snow is cold" is not the same as "snow is cold". I will use another example question. 1. John asked what time Sally came, and Ralph asked the same question. 2. The question of what time Sally came popped into Fred's mind, but was never asked. 3. The question of what time Sally came was answered incorrectly by Joe. 4. Mark determined the answer to what time Sally came by driving her in his car. 5. The question of what time Sally came was expressed in Spanish by a Jose. While #1 talks of speech acts, two askings of the same question, #2 involves no speech act. #5 shows that a question is not tied to the words used to expressed it. A question is a meaning, similar to a proposition, formed in the mind. Like a proposition, it can have multiple representations, but it is not a proposition. And we should not confuse a question with the acts of asking it. Enjoy, Don Subject: RE: [SBVR-FTF] issue 10621 -- Questions Date: Fri, 26 Jan 2007 11:12:27 -0700 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: [SBVR-FTF] issue 10621 -- Questions Thread-Index: AcdBXOViXdzi10DMTgOch14XvJ2qMgAGEhEg From: "Andy Carver" To: , X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id l0QJNaeu015090 Donald, Thanks for this. I'll look through it, but at the moment I'm suprised by your e-mail's equating of "illocutionary force" and "speech act". Two questions, for example, may have the same illocutionary force but be different speech acts (because of different locutionary meanings, e.g.). As I say, though, I'll look at these documents and see what they have to say. Thanks again, Andy -----Original Message----- From: Donald Chapin [mailto:Donald.Chapin@btinternet.com] Sent: Friday, January 26, 2007 8:11 AM To: sbvr-ftf@omg.org Subject: RE: [SBVR-FTF] issue 10621 -- Questions Ed, Don & Andy, The "performative (illocutionary force)| + proposition" construct has been a fundamental part of the SBVR architecture from the beginning (BRT decision at Sept. 2003 OMG Meeting in Boston). Attached is the latest working document (dec. 2003) that took a total SBVR architectural view, as well as a slide that summarizes "proposition + performative" (taken from the Boston meeting presentations). Dealing with proposition separately from illocutionary force and dealing with "proposition + illocutionary force" was a key reason for the UNCEFACT excitement about SBVR. Illocutionary force (speech acts) is a key element in the ISO TC 37 SC 4 language parsing standards which will need to be connectable with SBVR. While we clearly have not fully set forth the "performative + proposition" architecture in the current SBVR document (and should not try to do so before finalization), nothing that is in SBVR should contradict it. We should neither conflate nor confuse the constructs for 'proposition', 'performative (illocutionary force, speech act)' or 'proposition + performative'. Donald > -----Original Message----- > From: Ed Barkmeyer [mailto:edbark@nist.gov] > Sent: 26 January 2007 00:01 > To: Baisley, Donald E > Cc: sbvr-ftf@omg.org > Subject: Re: [SBVR-FTF] issue 10621 -- Questions > > Baisley, Donald E wrote: > > >>So if 'question' were independent of all speech acts, > >>it would be indistinguishable from 'proposition'. > > > > "whether snow is cold" means "what truth-value is had by the proposition > > that snow is cold". The meaning of "whether snow is cold" is not the > > same as "snow is cold". > > But in this sense, asserting "that snow is cold" is not the same as "snow > is > cold", either. The meaning of the assertion "that snow is cold" is "the > truth > value of the proposition 'snow is cold' is 'true'." > > This is the problem. A question is different from a proposition in that > it is > a speech act that operates on a proposition. An assertion is different > from a > proposition in that it is a speech act that operates on a proposition. > They > both differ in meaning from the proposition itself by the fact of the > speech > act. And they differ from each other in the nature of the speech act. > > Questions, assertions, denials, etc. are all speech acts that operate on > propositions. Those speech acts add meaning. The point of 10621 is that > the > taxonomy of 'meaning' might appropriately be 'concept', 'proposition', > 'speech > act' (I'm not sure of that), but 'question' is just one subtype of 'speech > act', and ALL speech acts are different from propositions and concepts. > Propositions involve concepts, and speech acts involve propositions. > > > I will use another example question. ... > > > ... A question is a meaning, similar to a > > proposition, formed in the mind. Like a proposition, it can have > > multiple representations, but it is not a proposition. And we should > > not confuse a question with the acts of asking it. > > Yes, a question is a meaning. I don't know enough about speech act theory > to > say, but I think that the term speech act refers to the meaning that is > added > to a proposition when the proposition is used in communication. It is not > the > physical act of asking a question that makes it a question. It is the > conceptual act of evaluating a proposition, as distinct from the act of > asserting the proposition or denying it or mentioning it. > > Put another way, when I was a debater in high school and college, the > stage > would be set with words like: Today we will debate the proposition > "Resolved, > that the United Nations has improved the African economies." Debating the > proposition means discussing whether the proposition holds. The side > called > the Affirmation asserts the proposition; the side called the Negation > denies > the proposition. Debating or discussing a proposition is similar to > "asking > the question". It is not clear that anyone "asks the question", but they > all > act on the proposition. The point is that there is nothing special about > 'question' here -- there are several different acts on the proposition, > and > posing the proposition as a 'question' is just one of them. > > Issue 10621 is in part whether 'question' is the parallel to 'proposition' > and > 'concept'. And I think it is, but it is only an example of meanings that > are > parallel to 'proposition'. An assertion is a meaning, a denial is a > meaning, > an imperative is a meaning. They are all different from proposition and > concept and from question. Where do they appear in the SBVR taxonomy? > How > does SBVR model them? As mentions? What is special about question? > > I also think SBVR has rolled up two somehow different ideas in the term > 'question': the evaluation of a proposition, and the enumeration of the > extension of a concept. They act on different things. I don't know where > this takes us. > > Issue 10621 says in effect that we have one term for two concepts, and one > of > those concepts -- the speech act -- is a proper subtype of a class of > meanings > we might want to include in SBVR. > > I wish I knew what theory should be applied to resolve this concern. I > don't. > > -Ed > > -- > Edward J. Barkmeyer Email: edbark@nist.gov > National Institute of Standards & Technology > Manufacturing Systems Integration Division > 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 > Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 > > "The opinions expressed above do not reflect consensus of NIST, > and have not been reviewed by any Government authority." Subject: RE: [SBVR-FTF] issue 10621 -- Questions Date: Fri, 26 Jan 2007 17:10:11 -0800 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: [SBVR-FTF] issue 10621 -- Questions Thread-Index: AcdBXOViXdzi10DMTgOch14XvJ2qMgAGEhEgAA0PnEA= From: "Baisley, Donald E" To: X-OriginalArrivalTime: 27 Jan 2007 01:10:12.0735 (UTC) FILETIME=[E569D8F0:01C741AF] X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id l0R12Sqa021080 My recollection from looking into "speech acts" early in the BSBR days was that speech acts involve "closed propositions" and "open propositions". Different acts are appropriate to these different kinds of "propositions". SBVR's "proposition" is a "closed proposition". SBVR's "question" is an "open proposition". I think we all agree that we can assume that the meaning of: Don's office is in California. is a proposition, and is the same proposition that is meant by: Le bureau de Don est en Californie. It should not be hard to agree then that the meaning of: In what state is Don's office? is a question, and is the same question that is meant by: Dans quel état est-il le bureau de Don ? Pardon my French, but I just want to point out that there is something we call a question in SBVR that is a meaning, and like other meanings, one question can have multiple representations. SBVR talks about speech acts in the introduction to chapter 8. Terry provided some important thoughts that are captured there. SBVR's 'question' is independent of speech acts in the sense that the same question can be involved in multiple different speech acts. In that way it is just like a proposition. Also, it can be represented by a text independent of any speech act. If we can all agree that: In what state is Don's office? represents a meaning, then we must agree that there is not a problem having SBVR's 'question' specialize 'meaning'. So if there is a problem to fix here, it is a matter of clarifying the definition of 'question' and not a problem requiring a change to the taxonomy of 'meaning'. This is important. It lowers the priority of the issue and prevents the need for a giant job that would result from changing SBVR at such a general point in its taxonomy. This is in chapter 8, something shared by both the Logical Formulation of Semantics vocabulary and the business facing vocabularies of chapters 11 and 12. If there needs to be more about this from a formal perspective, I don't have a problem with elaboration being added in chapter 10. Regards, Don DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=btinternet.com; h=Received:X-YMail-OSG:Reply-To:From:To:References:Subject:Date:Organization:Message-ID:MIME-Version:Content-Type:X-Mailer:X-MimeOLE:Thread-Index:In-Reply-To; b=OBecUazFQK1Ytar583goTk6cZ0LAhnIRv5zoCXG0DhJRAHQgpV2Dr1tX6hwq1WHKz3/Ucz4YMDB6G4qFviFZm+a4LT6ggJZBZbHM/WF7euIVjNMp6N8WC5q5d7cr0C604tofTT4Ysx0Ud3+EzBtiEmANJuWbEzPOTUAWAs4U3j0= ; X-YMail-OSG: hwM3oY8VM1kQ4x1CDPJv2O9UOevBCvutI85qVB2hA8OTGYz5UthU7SfOUmmtPWTQ.HaF0mztnocegG0f3gyWd_RJ0e3oprXWxWeuok_GaAyK7Y6OI3gAozvueznaX6hT.mZQq1q5qxIyenZCHQpPL3tJGA3hoeYrV1JXKRClEx3RRGM3ADznsgQ7kA-- Reply-To: From: "Donald Chapin" To: Subject: RE: issue 10621 -- SBVR FTF issue Date: Sat, 27 Jan 2007 11:17:35 -0000 Organization: Business Semantics Ltd X-Mailer: Microsoft Office Outlook 11 Thread-Index: Acc/NFA3L29cT60ITrStLPI3YwSUlAC0Bmsg While preparing Ballot 3, I was reminded that this Issue (10621) is part of the resolution of Issue 9715, and is referenced as such in it. This Issue (10621) therefore falls within the work of the 2nd SBVR FTF. Donald -------------------------------------------------------------------------------- From: Juergen Boldt [mailto:juergen@omg.org] Sent: 23 January 2007 21:20 To: issues@omg.org; sbvr-ftf@omg.org Subject: issue 10621 -- SBVR FTF issue This is issue # 10621 Source: Ed Barkmeyer, NIST, edbark@nist.gov Definition of 'question' In clause 8.1.3, 'question' is defined as "meaning of an interrogatory". Assuming that "interrogatory" means "interrogatory sentence", this defines the meaning in terms of the form of expression. Because the same interrogatory sentence may have more than one meaning, depending on the context of its utterance, this definition means: the meaning of an interrogatory sentence in context. But the same meaning extends across (some) contexts, while the answer may be different in each such context. That is, some part of the context of utterance affects the interpretation of the utterance; other parts of the context of utterance only affect its result. So this is a poor definition. A question is really an "operation" on a concept or proposition that creates a new form of meaning. One can describe a 'question' as a function applied to a concept or proposition whose result is its extension. There are 3 kinds of questions: - a concept question (What), which asks for the extension of a concept, - a proposition question (Whether), which asks for the truth-value (technically the extension) of a proposition. - a propositional relationship question (Why), which asks for the extension of a 2nd-order proposition about propositions (or objectified 'states of affairs' identified by propositions) -- the set of all propositions P such that 'P causes Q' or 'P entails Q'. Since clause 9 apparently supports 'question' meanings in exactly this way, SBVR should define 'question' to be a meaning that explicitly refers to the extension of a concept. That is, the question means its answer. It should not mean the action of asking. Juergen Boldt Director, Member Services Object Management Group 140 Kendrick St Building A Suite 300 Needham, MA 02494 USA tel: +1 781 444 0404 x 132 fax: +1 781 444 0320 email: juergen@omg.org www.omg.org Date: Mon, 29 Jan 2007 15:13:29 -0500 From: Ed Barkmeyer Reply-To: edbark@nist.gov Organization: NIST User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax) X-Accept-Language: en, fr, de, pdf, it, nl, sv, es, ru To: "Baisley, Donald E" CC: sbvr-ftf@omg.org Subject: Re: [SBVR-FTF] issue 10621 -- Questions X-MailScanner-Information: Please contact postmaster@mel.nist.gov for more information X-MailScanner: Found to be clean X-MailScanner-SpamCheck: X-MailScanner-From: edbark@nist.gov X-Spam-Status: No Baisley, Donald E wrote: My recollection from looking into "speech acts" early in the BSBR days was that speech acts involve "closed propositions" and "open propositions". Different acts are appropriate to these different kinds of "propositions". SBVR's "proposition" is a "closed proposition". SBVR's "question" is an "open proposition". I assume the following definitions: open proposition -- a logical formulation having the form of a (mathematical logic) 'proposition' involving at least one free variable. closed proposition -- a logical formulation having the form of a (mathematical logic) 'proposition' involving no free variables. In SBVR, unlike mathematical logic, we do *not* use the term 'proposition' to mean a logical formulation. And in SBVR we say that only a closed logical formulation formulates a 'proposition', meaning a 'meaning'. It is certainly possible to assign a "question" meaning to an open logical formulation, but in clause 9, we do not do that -- we rather assign it to a formulation of a concept, as a collection/set. So I don't see any relationship between these terms and the issue being discussed. I think we all agree that we can assume that the meaning of: Don's office is in California. is a proposition, and is the same proposition that is meant by: Le bureau de Don est en Californie. It should not be hard to agree then that the meaning of: In what state is Don's office? is a question, and is the same question that is meant by: Dans quel état est-il le bureau de Don ? Well, the two examples above involve the same predicate: x is-located-in y (1) Don's office is-located-in California (2) Don's office is-located-in ?state (1) can be interpreted as just a proposition, namely: (1p) "whether Don's office is-located-in California" (1) may also be interpreted as an assertion (fact): (1a) "Don's office is-located-in California" 'is true'. Those two interpretations are different meanings. The (unevaluated) proposition is distinct from the claim. So, no, I don't agree that we can all assume that the meaning of (1) is a proposition. In fact, most people would probably think it is an assertion. (2), according to SBVR, does not state a proposition, because it has a free variable. It does not ask for the truth value of a proposition, nor does it claim a truth value for a proposition. (2) may be said to correspond to a (possibly infinite) set of propositions, one for each possible value of ?state. Some (possibly none) of those propositions are true, some are false. And SBVR formulates the question as a 'projection' that in some sense includes all of those propositions (in its 'evaluation', but not in its 'result'). Now consider: (3) Is Don's office in California? (3) uses the same proposition as (1) = "whether Don's office is-located-in California" but (3) doesn't claim it to be true. Instead (3) asks for its truth value. That is, like (1a), (3) is not a proposition, but both (3) and (1a) involve the proposition (1p). Pardon my French, but I just want to point out that there is something we call a question in SBVR that is a meaning, and like other meanings, one question can have multiple representations. All of which is true, and utterly misses the two questions in 10621: How many distinct concepts have been rolled into the term 'question'? How is a question different from a proposition and a concept? The common characteristic is 'meaning'. What is the delimiting characteristic? If a question is the meaning of an interrogatory speech act, what is a proposition the meaning of? And is the meaning of an interrogatory speech act not a concept? It is actually easier for me to understand modeling a proposition as a kind of question -- the debate model. SBVR talks about speech acts in the introduction to chapter 8. Terry provided some important thoughts that are captured there. SBVR's 'question' is independent of speech acts in the sense that the same question can be involved in multiple different speech acts. From the discussion between Andy and Donald, I am coming to understand that what we are discussing is the "force" of a speech act. In that way it is just like a proposition. If I understand correctly, a proposition per se has no force. A question or assertion adds "illocutionary" force to a proposition. I may have this wrong, but I think the "illocutionary force" of an assertion is what SBVR calls a 'fact'. The force of an assertion is an evaluation of the proposition. The force of a (whether) question is a request for an evaluation of the proposition. If we can all agree that: In what state is Don's office? represents a meaning, then we must agree that there is not a problem having SBVR's 'question' specialize 'meaning'. I don't disagree with that at all. The problem is that we apparently define 'question' as the meaning of one kind of speech act, and we don't define any other kind of meaning as the meaning of any other kind of speech act. Neither a concept nor a proposition appears to be the meaning of any kind of speech act. Further, there is a relationship between one kind of question: "Is Don's office located in California?" and a corresponding proposition. And this kind of question does not seem to be related to the extension of a concept. But there is no relationship between the other kind of question: "What state is Don's office located in?" and any corresponding proposition. But this kind of question is related to the extension of a concept: the state ?s such that Don's office is-located-in ?s So if there is a problem to fix here, it is a matter of clarifying the definition of 'question' and not a problem requiring a change to the taxonomy of 'meaning'. This blandly supposes that we can clarify the definition without introducing missing concepts. This is important. It lowers the priority of the issue and prevents the need for a giant job that would result from changing SBVR at such a general point in its taxonomy. This is in chapter 8, something shared by both the Logical Formulation of Semantics vocabulary and the business facing vocabularies of chapters 11 and 12. Well, I thought this was important, because we should have a common understanding of meaning if we are going to subtype it. It seems to me that 'question' is a late addition to the taxonomy that wasn't very well thought out, and that lack of analysis shows up in the formulations of question and answer in clause 9 (as I said 15 months ago). And OBTW, 'question' does not show up in clause 11 or 12. But we don't have to address this. No one who implements SBVR will ever either understand or care. The high priority right now is to get this thing out and finalized, to minimize further cost. -Ed P.S. Thank you, Donald and Andy, for your contributions. I have learned something, and indeed something useful, from them. -- Edward J. Barkmeyer Email: edbark@nist.gov National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 FAX: +1 301-975-4694 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." User-Agent: Microsoft-Entourage/11.2.3.060209 Date: Wed, 31 Jan 2007 16:42:58 -1000 Subject: Re: issue 10621 -- SBVR FTF issue - Resolution Document From: keri To: SBVR-FTF Thread-Topic: issue 10621 -- SBVR FTF issue - Resolution Document Thread-Index: AcdFqq6i7QYBArGdEduT2gARJM+Cgg== Attached, per today.s meeting. - Keri On 1/23/07 11:20 AM, "Juergen Boldt" wrote: This is issue # 10621 Source: Ed Barkmeyer, NIST, edbark@nist.gov Definition of 'question' In clause 8.1.3, 'question' is defined as "meaning of an interrogatory". Assuming that "interrogatory" means "interrogatory sentence", this defines the meaning in terms of the form of expression. Because the same interrogatory sentence may have more than one meaning, depending on the context of its utterance, this definition means: the meaning of an interrogatory sentence in context. But the same meaning extends across (some) contexts, while the answer may be different in each such context. That is, some part of the context of utterance affects the interpretation of the utterance; other parts of the context of utterance only affect its result. So this is a poor definition. A question is really an "operation" on a concept or proposition that creates a new form of meaning. One can describe a 'question' as a function applied to a concept or proposition whose result is its extension. There are 3 kinds of questions: - a concept question (What), which asks for the extension of a concept, - a proposition question (Whether), which asks for the truth-value (technically the extension) of a proposition. - a propositional relationship question (Why), which asks for the extension of a 2nd-order proposition about propositions (or objectified 'states of affairs' identified by propositions) -- the set of all propositions P such that 'P causes Q' or 'P entails Q'. Since clause 9 apparently supports 'question' meanings in exactly this way, SBVR should define 'question' to be a meaning that explicitly refers to the extension of a concept. That is, the question means its answer. It should not mean the action of asking. Juergen Boldt Director, Member Services Object Management Group 140 Kendrick St Building A Suite 300 Needham, MA 02494 USA tel: +1 781 444 0404 x 132 fax: +1 781 444 0320 email: juergen@omg.org www.omg.org Issue10621.doc Disposition: Deferred OMG Issue No: 10621 Title: Definition of 'question' Source: Ed Barkmeyer, NIST Summary: In clause 8.1.3, 'question' is defined as "meaning of an interrogatory". Assuming that "interrogatory" means "interrogatory sentence", this defines the meaning in terms of the form of expression. Because the same interrogatory sentence may have more than one meaning, depending on the context of its utterance, this definition means: the meaning of an interrogatory sentence in context. But the same meaning extends across (some) contexts, while the answer may be different in each such context. That is, some part of the context of utterance affects the interpretation of the utterance; other parts of the context of utterance only affect its result. So this is a poor definition. A question is really an "operation" on a concept or proposition that creates a new form of meaning. One can describe a 'question' as a function applied to a concept or proposition whose result is its extension. There are 3 kinds of questions: - a concept question (What), which asks for the extension of a concept, - a proposition question (Whether), which asks for the truth-value (technically the extension) of a proposition. - a propositional relationship question (Why), which asks for the extension of a 2nd-order proposition about propositions (or objectified 'states of affairs' identified by propositions) -- the set of all propositions P such that 'P causes Q' or 'P entails Q'. Since clause 9 apparently supports 'question' meanings in exactly this way, SBVR should define 'question' to be a meaning that explicitly refers to the extension of a concept. That is, the question means its answer. It should not mean the action of asking. Resolution: The FTF discussed this issue at length and could not agree on a resolution. Disposition: Deferred DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=btinternet.com; h=Received:X-YMail-OSG:Reply-To:From:To:Subject:Date:Organization:Message-ID:MIME-Version:Content-Type:X-Mailer:Thread-Index:X-MimeOLE; b=PAFT9dH63nZ1bSDn2o9h6srtHoSFIat2HApAuiec6exyrKSKd7zwYOGfIh5vPYfwQn+dOKWeTWl5jFYoJfWaA+sMsB6aqKTIqd8K5EEhcZ853r0BJda6dJtj/1X/Z1jdF604h23yL00AcyeRpnEepgrhCvgrZbAbquZjx5qcNZ8= ; X-YMail-OSG: iok7macVM1naESQyb5pwj_30y0wAdOYj5vXYxg.q.pYqC_R8npw4EH0pnSjeLCD4fpEIWmMrbdRQOq3xqFloQ7jc5Lit8aGATeoUF_Ros5_WoIfdtoxgfCMDlFa7lfGXGJ0.EOWpnAzKUAs52lgpvtD5yOl1ok1wWCBnjm3ImPzQUTk8siOot.8iuw-- Reply-To: From: "Donald Chapin" To: Subject: FW: issue 10621 -- SBVR FTF issue - Resolution Document Date: Wed, 7 Feb 2007 13:25:18 -0000 Organization: Business Semantics Ltd X-Mailer: Microsoft Office Outlook 11 Thread-Index: AcdFqq6i7QYBArGdEduT2gARJM+CggFD96FA I have added some wording to the Deferral of Issue 10621 that I heard in the telephone meeting that released this Issue to Ballot. It is important that the SBVR FTF report is very clear up front as to the impact on the SBVR Specification of deferring an Issue to the SBVR FTF. Clear statements of minimal impact on the quality of the SBVR Available Specification will minimize problems with the Architecture Board approval. Donald -------------------------------------------------------------------------------- From: keri [mailto:keri_ah@mac.com] Sent: 01 February 2007 02:43 To: SBVR-FTF Subject: Re: issue 10621 -- SBVR FTF issue - Resolution Document Attached, per today.s meeting. - Keri On 1/23/07 11:20 AM, "Juergen Boldt" wrote: This is issue # 10621 Source: Ed Barkmeyer, NIST, edbark@nist.gov Definition of 'question' In clause 8.1.3, 'question' is defined as "meaning of an interrogatory". Assuming that "interrogatory" means "interrogatory sentence", this defines the meaning in terms of the form of expression. Because the same interrogatory sentence may have more than one meaning, depending on the context of its utterance, this definition means: the meaning of an interrogatory sentence in context. But the same meaning extends across (some) contexts, while the answer may be different in each such context. That is, some part of the context of utterance affects the interpretation of the utterance; other parts of the context of utterance only affect its result. So this is a poor definition. A question is really an "operation" on a concept or proposition that creates a new form of meaning. One can describe a 'question' as a function applied to a concept or proposition whose result is its extension. There are 3 kinds of questions: - a concept question (What), which asks for the extension of a concept, - a proposition question (Whether), which asks for the truth-value (technically the extension) of a proposition. - a propositional relationship question (Why), which asks for the extension of a 2nd-order proposition about propositions (or objectified 'states of affairs' identified by propositions) -- the set of all propositions P such that 'P causes Q' or 'P entails Q'. Since clause 9 apparently supports 'question' meanings in exactly this way, SBVR should define 'question' to be a meaning that explicitly refers to the extension of a concept. That is, the question means its answer. It should not mean the action of asking. Juergen Boldt Director, Member Services Object Management Group 140 Kendrick St Building A Suite 300 Needham, MA 02494 USA tel: +1 781 444 0404 x 132 fax: +1 781 444 0320 email: juergen@omg.org www.omg.org B Date: Thu, 17 Mar 2011 13:05:36 -0400 From: Ed Barkmeyer Reply-To: edbark@nist.gov Organization: NIST User-Agent: Thunderbird 2.0.0.23 (Windows/20090812) To: SBVR RTF Subject: Draft Resolution of SBVR issue 10621 X-NISTMEL-MailScanner-Information: Please contact postmaster@mel.nist.gov for more information X-NISTMEL-MailScanner-ID: p2HH5fYo001376 X-NISTMEL-MailScanner: Found to be clean X-NISTMEL-MailScanner-SpamCheck: X-NISTMEL-MailScanner-From: edbark@nist.gov X-NISTMEL-MailScanner-Watermark: 1300986342.16677@oQ2/ORaR2tHVwgQ2mAjJFA X-Spam-Status: No X-NIST-MailScanner: Found to be clean X-NIST-MailScanner-From: edbark@nist.gov All, The draft resolution to Issue 10621, from the OMG meeting in June 2010, is attached. Note: The meeting minutes also mentioned adding 'whether' to Annex C, in order to allow the meaning that is modified by a question to be a proposition. But there does not seem to be a way to formulate that in clause 9, because the meaning associated with a question or answer nominalization is represented by a projection, rather than a logical formulation. So I don't see a reason to create the capability in Structured English to express a meaning for which we define no standard exchange form. Note also that the change in clause 8 causes only a small wording change in clause 9, because it doesn't change the intent. The added fact type in clause 8 is there solely to give us the terminology to describe the concept. No 'question' will ever appear as an independent meaning in a vocabulary or conceptual schema; the concept is there because questions appear within rules. -Ed -- Edward J. Barkmeyer Email: edbark@nist.gov National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 Cel: +1 240-672-5800 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." SBVR Issue 10621-d1.doc Disposition: Resolved OMG Issue No: 10621 Title: Definition of 'question' Source: Ed Barkmeyer, NIST, edbark@nist.gov Summary: In clause 8.1.3, 'question' is defined as "meaning of an interrogatory". Assuming that "interrogatory" means "interrogatory sentence", this defines the meaning in terms of the form of expression. Because the same interrogatory sentence may have more than one meaning, depending on the context of its utterance, this definition means: the meaning of an interrogatory sentence in context. But the same meaning extends across (some) contexts, while the answer may be different in each such context. That is, some part of the context of utterance affects the interpretation of the utterance; other parts of the context of utterance only affect its result. So this is a poor definition. A question is really an "operation" on a concept or proposition that creates a new form of meaning. One can describe a 'question' as a function applied to a concept or proposition whose result is its extension. There are 3 kinds of questions: - a concept question (What), which asks for the extension of a concept, - a proposition question (Whether), which asks for the truth-value (technically the extension) of a proposition. - a propositional relationship question (Why), which asks for the extension of a 2nd-order proposition about propositions (or objectified 'states of affairs' identified by propositions) -- the set of all propositions P such that 'P causes Q' or 'P entails Q'. Since clause 9 apparently supports 'question' meanings in exactly this way, SBVR should define 'question' to be a meaning that explicitly refers to the extension of a concept. That is, the question means its answer. It should not mean the action of asking. Resolution: The RTF agrees that a 'question' adds interrogatory meaning a concept or proposition, and should be defined to do that. And therefore, a fact type will be added to relate the 'question' to the concept or proposition to which it adds meaning. 'Ask' and 'answer' are different verbs associated with a 'question'. Minor changes to the wording of clause 9 for question and answer nominalization will also be made. Revised Text: 1a. In clause 8.1.3 in the entry for 'question', REPLACE the Definition: Definition: meaning of an interrogatory with: Definition: meaning that adds an interrogatory force to another meaning (a concept or a proposition) 1b. In clause 8.1.3 in the entry for 'question', REPLACE the Reference Scheme Reference Scheme: a closed projection that means the question with the following two paragraphs: Reference Scheme: the meaning that the question modifies Example: 'The clerk must ask the customer what kind of car the customer wants' includes a 'question': 'what kind of car the customer wants', which modifies the concept: 'the kind of car that the customer wants'. 2. In clause 8.1.3 after the entry for 'question', ADD a new entry: question modifies meaning Definition: the question adds an interrogatory force to the meaning 3. In clause 9.2.9, in the entry for 'question nominalization', in the Definition paragraph, CHANGE the text "question that is meant by the projection" to the text: "question that modifies the concept that is meant by the projection" Definition: projecting formulation that formulates the meaning: the thing to which the bindable target that is bound to the projecting formulation refers is the question that modifies the concept that is meant by the projection of the projecting formulation 4. In clause 9.2.9, in the entry for 'answer nominalization', REPLACE the Definition paragraph: Definition: projecting formulation that formulates the meaning: the thing to which the bindable target that is bound to the projecting formulation refers is a proposition that is true and that completely and correctly answers the question meant by the projection of the projecting formulation with the two paragraphs: Definition: projecting formulation that formulates the meaning: the thing to which the bindable target that is bound to the projecting formulation refers is a proposition that a specific set of things is the extension of the concept that is meant by the projection of the projecting formulation, and further, that that proposition is true. Note: An answer formulation does not assign the bindable target to a set (the extension of the concept meant by the projection), but rather to a proposition that identifies a specific set of things as that extension. That is, it binds to a proposition that correctly answers the question that modifies the concept meant by the projection. Disposition: Resolved B B Date: Fri, 17 Jun 2011 18:52:17 -0400 From: Ed Barkmeyer Reply-To: edbark@nist.gov Organization: NIST User-Agent: Thunderbird 2.0.0.23 (Windows/20090812) To: SBVR RTF Subject: Issue 10621 issue expansion: text of 'closed projection means question' X-NISTMEL-MailScanner-Information: Please contact postmaster@mel.nist.gov for more information X-NISTMEL-MailScanner-ID: p5HMqLMJ028142 X-NISTMEL-MailScanner: Found to be clean X-NISTMEL-MailScanner-SpamCheck: X-NISTMEL-MailScanner-From: edbark@nist.gov X-NISTMEL-MailScanner-Watermark: 1308955944.54638@eyJWh1fE6pAgLyCU6wupxw X-Spam-Status: No X-NIST-MailScanner: Found to be clean X-NIST-MailScanner-From: edbark@nist.gov In the RTF telecon of 18 Mar 2011, discussion of Issue 10621 pointed to the fact that the question forms described in 'closed projection means formulation' (at the end of 9.3) are not described in Annex C, and should be added. 'What' is described in Annex C, and I don't think the others should be, although one could make a case for 'whether'. The actual text of the entry in 9.3 is: _closed projection_ /means/ _question _ Definition: the _closed projection_ formulates the _question_ such that the result of the projection answers the _question_ Necessity: Each closed projection means at most one question. Note: A question using an interrogative operator such as .what., .when., .where., .why., or .how. is generally formulated by a projection on a variable that ranges over a concept that matches the operator. The interrogative .what. is often used with a designation of a noun concept such as in .What car is available?. in which case the variable ranges over the noun concept .car.. For each of the other operators the variable ranges over a noun concept fitting to that operator as if .what. had been used with a designation for that concept. Examples of the correspondence of interrogative operators to noun concepts is shown below. .When is a car available?. What time .How is a car driven?. What method .Where is a car?. What location .Who can drive a car?. What person .Why is a car available?. What cause Note that definition of these nouns (underlined above) is outside the scope of SBVR. However, the concept .cause. is a role that ranges over the concept .state of affairs., so an answer to a .why. question is often formulated using objectification (the last example under objectification considers one state of affairs as a cause of another). Note: A true/false question is typically nominalized using the interrogative operator .whether. as in .The customer asked whether a car is available,. but is asked (in English) with no such operator: .Is a car available?.. The meaning of .whether. in this context is .What truth-value does this proposition have?.. The formulation of such a question is a projection on a variable that ranges over a characteristic type (here called .truth-value.) whose instances are the characteristics .proposition is true. and .proposition is false.. The projection is constrained by the truth-value being that of the proposition .a car is available. formulated using proposition nominalization. Example: .Is a car available?. The question is meant by a closed projection. . The projection is on a unitary variable. . . The variable ranges over the concept .truth-value.. ... Problem (1): If the Note that the nouns 'time', 'method', 'location', 'person', 'cause' are outside the scope of SBVR is correct, then SBVR lacks the semantic basis for using the terms 'when', 'how', 'where', 'who' in Annex C. This Note is more about guidance for toolsmiths, who may have such words in their grammars. Problem (2): 'who' and 'what' refer to instances of a concept that can be effectively described by a projection involving the instance variable in a characteristic role. That is, 'the clerk must ask 'what kind of car the customer wants' is a direct reference to instances of 'kind of car that the customer wants'. Similarly, 'the investigator shall determine who was at fault' makes direct reference to the concept 'person/thing that was at fault'. By comparison, 'when', 'where', 'how' and 'why' require the existence of concepts like time and location, and also require the existence of some fact type that is implicitly used to capture the intended relationship between the instances of the projection and the characterization of the state of affairs used in formulating the question. That is, 'when is a car available' requires the use of 'state of affairs occurs at time' or perhaps 'state of affairs begins at time' or 'state of affairs occurs within time'. So SBVR is well short of supplying the semantics for interpreting such a question. Indeed, each business application may choose an importantly different interpretation of 'when' relative to the usage, and would be better advised to formulate the question using a specific fact type relationship to 'time', or to 'location', etc. Problem (3): Note that 'what kind of car the customer wants' is an adjectival use of 'what', while 'what the customer wants' is a nominal (noun-like) use of 'what'. Annex C is perhaps not clear as to which of those is the intended meaning of 'what', but the intent is that it is used as an adjective. (It cannot mean both. Otherwise, what, for example, is the meaning of "what children sing"?) Problem (4): In the example use of 'whether', the projection is said to have a variable that ranges over proposition and the constraining logical formulation is 'proposition is true'. It follows that the meaning of the question is a set of (at most 1) proposition, rather than a truth value. And, as I read the definition of 'answer formulation', it would require that proposition to be used in another proposition, presumably of the form 'proposition is true'. In short, the formulation of the question 'whether a car is available' would require something other than a projection, or a special interpretation of a projection. In any case, the example in the text represents a contorted interpretation of 'closed projection means question', and is therefore a bad example. Finally, I was asked to modify the writeup to support questions like "Which golfers won the U.S. Open in which years?" -- questions related to projections in which there are two or more variables. The question is: What is the meaning of (Sammy Snead, 1956), (Jack Nicklaus, 1972)? This is the result of a database query -- a set of ordered pairs, onto which the user _imposes_ the semantics that answers the question. Of itself, the projection is just a set of meaningless pairs. When a projection is on a single variable, the result is consistently construed as the extent of the noun concept that is meant by the projecting formulation. When the projection is used to define a fact type, the projecting formulation is the definition of the fact type in terms of the role variables. But the extension of the fact type is a set of actualities, not a set of tuples of things. One can consider an actuality to be described by a tuple of things-in-roles (which is Terry's model in ORM), where the roles are unique to the fact type. And SBVR (now) defines the extension of a fact type in that way. So the interpretation of the question: "Which golfers won the U.S. Open in which years?" first creates an ad hoc fact type: (golfer) wins U.S.Open in (year) then identifies the instances of that fact type by: (golfer= Sammy Snead, year= 1956) and so on. But it still leaves open the question: Is the meaning of the question the set of actualities, or the set of whatever these pairs are? I do not believe that questions with multiple query variables EVER participate in business rules. And SBVR is not about restricted natural language for database queries. So I respectfully suggest that we just don't open that can of worms. All in all, I believe SBVR has tried to standardize a non-solution to less-than-adequately-researched problem. Further, as we recently observed, the relationship between questions and noun concept nominalizations is much closer than was previously recognized. Verbs of specification and change nominalize the direct object concept, in much the same way the question does. "The delivery date must not change" means "What the delivery date is must not change." And this in turn follows from the observation that led to this issue -- the a question adds an interrogatory force to another meaning -- a concept or a proposition. Sorry, Donald, I don't think this will be entirely resolved by Thursday. I attach two versions of the writeup, one with the two agreed-upon corrections from March, and one that addresses the issues above in a way that I'm sure no one will like. -Ed -- Edward J. Barkmeyer Email: edbark@nist.gov National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 Cel: +1 240-672-5800 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." SBVR Issue 10621-d2.doc From: Don Baisley To: SBVR RTF Subject: RE: Issue 10621 issue expansion: text of 'closed projection means question' Thread-Topic: Issue 10621 issue expansion: text of 'closed projection means question' Thread-Index: AQHMLUF7vcEr610N3EKOA7EU64nzIpTKSbrQ Date: Thu, 23 Jun 2011 03:48:34 +0000 Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [157.54.51.37] X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id p5N3fDNt019399 I am sympathetic to Ed wanting to see "question" defined in terms of other meanings. I also tried to find some basis in papers on formal semantics of questions that would support Ed's approach to defining questions in terms of concepts. I found a lot on the formal semantics of questions, but I did not find support for Ed's approach. Besides Ed's approach being novel rather than mainstream, it has a serious problem in that it breaks an important SBVR capability: questions with multiple variables (e.g., what renter is renting what EU-Rent vehicle?). I saw questions with multiple variables used in several paid proof-of-concept engagements dealing with customers' actual business rules and facts. Given the problem with Ed's approach, I prefer to leave the definition as it is. The definition is correct, even for the case Ed complained about, which is where an interrogatory has two meanings. That would happen if the same interrogatory sentence is interpreted by different speech communities. However, given that both communities see it as an interrogatory, both meanings are questions. Since SBVR is about semantics of business discourse, I don't have a problem defining "question" in terms of an element of discourse. Also, Ed's use of the signifier "interrogates" is a problem because that signifier already has a meaning in the subject area of "questions" that is different from Ed's novel meaning. If we are trying to add more clarity to the section on "question", I would add a fact type: '[proposition] answers [question]'. I would then describe what it takes for a proposition to completely answer a question. This can be determined logically and independently of whether the proposition is true. Such an addition to SBVR might be considered scope creep and outside of the RTF's charter. However, there is important clarity that comes from understanding that the answer to a question is a proposition, not an extension of some concept. Regards, Don -----Original Message----- From: Ed Barkmeyer [mailto:edbark@nist.gov] Sent: Friday, June 17, 2011 3:52 PM To: SBVR RTF Subject: Issue 10621 issue expansion: text of 'closed projection means question' In the RTF telecon of 18 Mar 2011, discussion of Issue 10621 pointed to the fact that the question forms described in 'closed projection means formulation' (at the end of 9.3) are not described in Annex C, and should be added. 'What' is described in Annex C, and I don't think the others should be, although one could make a case for 'whether'. The actual text of the entry in 9.3 is: > _closed projection_ /means/ _question > _ > Definition: the _closed projection_ formulates the _question_ such > that the result of the projection answers the _question_ > Necessity: Each closed projection means at most one question. > > Note: A question using an interrogative operator such as 'what', > 'when', 'where', 'why', or 'how' is generally formulated by a > projection on a variable that ranges over a concept that matches the > operator. The interrogative 'what' is often used with a designation of > a noun concept such as in "What car is available?" in which case the > variable ranges over the noun concept 'car'. For each of the other > operators the variable ranges over a noun concept fitting to that > operator as if 'what' had been used with a designation for that > concept. > Examples of the correspondence of interrogative operators to noun > concepts is shown below. > "When is a car available?" What time > "How is a car driven?" What method > "Where is a car?" What location > "Who can drive a car?" What person > "Why is a car available?" What cause > Note that definition of these nouns (underlined above) is outside the > scope of SBVR. > However, the concept 'cause' is a role that ranges over the concept > 'state of affairs', so an answer to a 'why' question is often > formulated using objectification (the last example under > objectification considers one state of affairs as a cause of another). > > Note: A true/false question is typically nominalized using the > interrogative operator 'whether' as in "The customer asked whether a > car is available," but is asked (in > English) with no such > operator: "Is a car available?". The meaning of 'whether' in this > context is "What truth-value does this proposition have?". The > formulation of such a question is a projection on a variable that > ranges over a characteristic type (here called 'truth-value') whose > instances are the characteristics 'proposition is true' and > 'proposition is false'. The projection is constrained by the > truth-value being that of the proposition "a car is available" > formulated using > proposition nominalization. > > Example: "Is a car available?" > The question is meant by a closed projection. > . The projection is on a unitary variable. > . . The variable ranges over the concept 'truth-value'. > ... Problem (1): If the Note that the nouns 'time', 'method', 'location', 'person', 'cause' are outside the scope of SBVR is correct, then SBVR lacks the semantic basis for using the terms 'when', 'how', 'where', 'who' in Annex C. This Note is more about guidance for toolsmiths, who may have such words in their grammars. Problem (2): 'who' and 'what' refer to instances of a concept that can be effectively described by a projection involving the instance variable in a characteristic role. That is, 'the clerk must ask 'what kind of car the customer wants' is a direct reference to instances of 'kind of car that the customer wants'. Similarly, 'the investigator shall determine who was at fault' makes direct reference to the concept 'person/thing that was at fault'. By comparison, 'when', 'where', 'how' and 'why' require the existence of concepts like time and location, and also require the existence of some fact type that is implicitly used to capture the intended relationship between the instances of the projection and the characterization of the state of affairs used in formulating the question. That is, 'when is a car available' requires the use of 'state of affairs occurs at time' or perhaps 'state of affairs begins at time' or 'state of affairs occurs within time'. So SBVR is well short of supplying the semantics for interpreting such a question. Indeed, each business application may choose an importantly different interpretation of 'when' relative to the usage, and would be better advised to formulate the question using a specific fact type relationship to 'time', or to 'location', etc. Problem (3): Note that 'what kind of car the customer wants' is an adjectival use of 'what', while 'what the customer wants' is a nominal (noun-like) use of 'what'. Annex C is perhaps not clear as to which of those is the intended meaning of 'what', but the intent is that it is used as an adjective. (It cannot mean both. Otherwise, what, for example, is the meaning of "what children sing"?) Problem (4): In the example use of 'whether', the projection is said to have a variable that ranges over proposition and the constraining logical formulation is 'proposition is true'. It follows that the meaning of the question is a set of (at most 1) proposition, rather than a truth value. And, as I read the definition of 'answer formulation', it would require that proposition to be used in another proposition, presumably of the form 'proposition is true'. In short, the formulation of the question 'whether a car is available' would require something other than a projection, or a special interpretation of a projection. In any case, the example in the text represents a contorted interpretation of 'closed projection means question', and is therefore a bad example. Finally, I was asked to modify the writeup to support questions like "Which golfers won the U.S. Open in which years?" -- questions related to projections in which there are two or more variables. The question is: What is the meaning of (Sammy Snead, 1956), (Jack Nicklaus, 1972)? This is the result of a database query -- a set of ordered pairs, onto which the user _imposes_ the semantics that answers the question. Of itself, the projection is just a set of meaningless pairs. When a projection is on a single variable, the result is consistently construed as the extent of the noun concept that is meant by the projecting formulation. When the projection is used to define a fact type, the projecting formulation is the definition of the fact type in terms of the role variables. But the extension of the fact type is a set of actualities, not a set of tuples of things. One can consider an actuality to be described by a tuple of things-in-roles (which is Terry's model in ORM), where the roles are unique to the fact type. And SBVR (now) defines the extension of a fact type in that way. So the interpretation of the question: "Which golfers won the U.S. Open in which years?" first creates an ad hoc fact type: (golfer) wins U.S.Open in (year) then identifies the instances of that fact type by: (golfer= Sammy Snead, year= 1956) and so on. But it still leaves open the question: Is the meaning of the question the set of actualities, or the set of whatever these pairs are? I do not believe that questions with multiple query variables EVER participate in business rules. And SBVR is not about restricted natural language for database queries. So I respectfully suggest that we just don't open that can of worms. All in all, I believe SBVR has tried to standardize a non-solution to less-than-adequately-researched problem. Further, as we recently observed, the relationship between questions and noun concept nominalizations is much closer than was previously recognized. Verbs of specification and change nominalize the direct object concept, in much the same way the question does. "The delivery date must not change" means "What the delivery date is must not change." And this in turn follows from the observation that led to this issue -- the a question adds an interrogatory force to another meaning -- a concept or a proposition. Sorry, Donald, I don't think this will be entirely resolved by Thursday. I attach two versions of the writeup, one with the two agreed-upon corrections from March, and one that addresses the issues above in a way that I'm sure no one will like. -Ed -- Edward J. Barkmeyer Email: edbark@nist.gov National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 Cel: +1 240-672-5800 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." From: Don Baisley To: "sbvr-rtf@omg.org" Subject: RE: Issue 10621 issue expansion: text of 'closed projection means question' Thread-Topic: RE: Issue 10621 issue expansion: text of 'closed projection means question' Thread-Index: Acw0VhFL+wLS++qpTEiZTHE/Fj7WEw== Date: Sun, 26 Jun 2011 23:09:14 +0000 Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [157.54.51.35] In last week.s SBVR meeting I was asked to prepare an example of a question with multiple variables and to prepare an example of an answer. The examples below assume numeric references schemes for cars and renters. Question: At the EU-Rent Redmond branch on June 27, 2011, which cars are assigned to which renters? Answer: At the EU-Rent Redmond branch on June 27, 2011, the assignings of cars to renters are the assigning of car #91 to renter #123 and the assigning of car #92 to renter #345. Note that the answer is definite, not open. This is the same as for a question with a single variable, such as this one: Question: Which cars are assigned to some renter at the EU-Rent Redmond branch on June 27, 2011? Proper (definite) Answer: The cars that are assigned to some renter at the EU-Rent Redmond branch on June 27, 2011, are car #91 and car #92. Improper (indefinite) Answer: At the EU-Rent Redmond branch on June 27, 2011, car #91 and car#92 are assigned to some renter. Formulation of a definite, nonsingular, answer to either question uses an aggregation formulation, which has a projection. In a multivariable case, the projection.s variable refers to assignings of cars to renters. Suppose assignings of cars to renters could be identified by some other reference scheme, such as by .assignment number.. It would be incorrect to answer the first question using assignment numbers rather than by referring specifically to individual cars and renters. This is because the question asks for the involved cars and renters specifically. Regards, Don Date: Tue, 28 Jun 2011 13:40:52 -0400 From: Ed Barkmeyer Reply-To: edbark@nist.gov Organization: NIST User-Agent: Thunderbird 2.0.0.23 (Windows/20090812) To: Don Baisley CC: "sbvr-rtf@omg.org" Subject: Re: Issue 10621 issue expansion: text of 'closed projection means question' X-NISTMEL-MailScanner-Information: Please contact postmaster@mel.nist.gov for more information X-NISTMEL-MailScanner-ID: p5SHeuIf013421 X-NISTMEL-MailScanner: Found to be clean X-NISTMEL-MailScanner-SpamCheck: X-NISTMEL-MailScanner-From: edbark@nist.gov X-NISTMEL-MailScanner-Watermark: 1309887659.60978@YuVIB6eBrzy4d4AOhKWH+Q X-Spam-Status: No X-NIST-MailScanner: Found to be clean X-NIST-MailScanner-From: edbark@nist.gov Don Baisley wrote: In last week.s SBVR meeting I was asked to prepare an example of a question with multiple variables and to prepare an example of an answer. The examples below assume numeric references schemes for cars and renters. Question: At the EU-Rent Redmond branch on June 27, 2011, which cars are assigned to which renters? Answer: At the EU-Rent Redmond branch on June 27, 2011, the assignings of cars to renters are the assigning of car #91 to renter #123 and the assigning of car #92 to renter #345. Notice that this is not about phrasing a fact or a business rule, but rather about phrasing a database query. It is not a requirement that an SBVR question have a direct mapping to SQL. More on this below. Note that the answer is definite, not open. Note that the phrasing of the answer that Don uses makes the result of the query a set of actualities, each stated as a gerund involving the queried elements in roles: "the assigning of car #91 to renter #123". If that set of actualities is the meaning of the projection, it is not at all what clause 9 says. This is the same as for a question with a single variable, such as this one: Question: Which cars are assigned to some renter at the EU-Rent Redmond branch on June 27, 2011? Proper (definite) Answer: The cars that are assigned to some renter at the EU-Rent Redmond branch on June 27, 2011, are car #91 and car #92. Right, but this is not at all the same as the above. The answer to this query is a set of cars -- a set of instances of the concept following 'what' in the question, i..e, a set of instances of the concept 'cars that are assigned to some rental ...'. Further, the question construct means 'the extension of that concept'. When used within a fact or a rule it 'nominalizes' the extension, as distinct from nominalizing the concept. But if the interpretation were the same as the first question, what would be returned is a set of actualities, i.e., 'car#91 being assigned to some rental', and 'car#92 being assigned to some rental', not 'car#91' and 'car#92'. Don cannot state a consistent semantics for the two queries. He adjusts the semantics of the operation to get the answer HE means, not the answer IT means. What this suggests is that there are two constructs -- single variable projections and multi-variable projections -- and they have very different semantics. Improper (indefinite) Answer: At the EU-Rent Redmond branch on June 27, 2011, car #91 and car#92 are assigned to some renter. Agreed. This does not state the extension of a concept. It simply identifies two instances. But this is not the issue at hand. Formulation of a definite, nonsingular, answer to either question uses an aggregation formulation, which has a projection. In a multivariable case, the projection.s variable refers to assignings of cars to renters. But in the single variable case, it doesn't? So, should we write a rule that says it means the set of actualities if there are multiple variables? Suppose assignings of cars to renters could be identified by some other reference scheme, such as by .assignment number.. It would be incorrect to answer the first question using assignment numbers rather than by referring specifically to individual cars and renters. This is because the question asks for the involved cars and renters specifically. So the answer is not the set of actualities, either. It is required to be a particular set of representations of those actualities that use the fact type form that was used in the question. It seems that we have come a long way from this formulation being a representation of an interrogatory force added to a concept. Don wants to require a particular form for representation of the extension. Or perhaps that is only a required property of an answer formulation. This seems to be entirely out of place. The question formulation nominalizes the extension -- it makes the extension a set that we can talk about, without naming its actual members. So the representation of the members is not part of its semantics. What is important is what concept it is a set of. If I understand Don's position correctly: - if there is only one projection variable, the extension is the set of things that the variable ranges over and that satisfy the proposition meant by the logical formulation. - if there is more than one projection variable, the extension is the set of actualities that make the propositions resulting from substitution of things into the variable slots in the logical formulation true. Concerns about the form of the statements used to formulate the answer in practice are entirely out of scope. Let us consider a business rule that is not a database query: "The racing form shall specify the horses and their pole positions in each race." could be written: "The racing form shall specify what horse occupies which pole position in each race." Or it could be written: "The racing form shall specify what horse occupies each pole position in each race." Unlike the two-variable form above ("what horse occupies which pole position"), this latter form has only one projection variable ("what horse"), but it explicitly binds 'race' and 'pole position' (which appear in the logical formulation of the projection) with universal quantifiers outside the projection. This has the effect of requiring the racing form to specify the answers to a set of questions. From a business semantics point-of-view, there is no difference in the required information content of the racing form, and we are not specifying how it is to be presented. So let us go back to Don's original query: At the EU-Rent Redmond branch on June 27, 2011, which cars are assigned to which renters? and make a related business rule: The daily log shall specify which cars are assigned to which renters on the day of the log. But it could be written: The daily log shall specify which car was assigned to each renter on the day of the log. E presto! Same semantics, single projection variable. The required information content is the answers to set of questions, one per renter. The answers will perforce associate renters with cars, because each question is distinguished by the renter involved. So, we can overcome semantic ambiguity and special interpretation rules in SBVR by forcing the business rule to be more clearly formulated. If the tool wants to accept the first rule above, that is a human-computer-interface issue. The second form, however, is what the tool should understand the first rule to mean, and it has a clear mapping to SBVR clause 9 and a clear interpretation for any receiving tool. My original broad statement that there are no two-variable questions in rules was clearly false. The issue is whether there _need to be_ two-variable questions in rules, and whether SBVR needs to support such questions directly. It is now my position (thanks to this discussion) that such rules can be restated as a quantified set of questions, each of which involves a single variable. And I think that approach actually has clearer semantics. -Ed Regards, Don -- Edward J. Barkmeyer Email: edbark@nist.gov National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 Cel: +1 240-672-5800 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." Date: Wed, 29 Jun 2011 11:01:17 -0400 From: Ed Barkmeyer Reply-To: edbark@nist.gov Organization: NIST User-Agent: Thunderbird 2.0.0.23 (Windows/20090812) To: SBVR RTF Subject: SBVR Issue 10621 draft 5 X-NISTMEL-MailScanner-Information: Please contact postmaster@mel.nist.gov for more information X-NISTMEL-MailScanner-ID: p5TF1MgN021451 X-NISTMEL-MailScanner: Found to be clean X-NISTMEL-MailScanner-SpamCheck: X-NISTMEL-MailScanner-From: edbark@nist.gov X-NISTMEL-MailScanner-Watermark: 1309964484.00727@SwssQpXo7bH7poBEu0KEoQ X-Spam-Status: No X-NIST-MailScanner: Found to be clean X-NIST-MailScanner-From: edbark@nist.gov I attach a revised version of the resolution of Issue 10621. As directed, I added an 'interrogatory' concept to clause 8.3. After careful review of the problem of 'whether', I concluded that the least change resolution is to simply justify the use of 'proposition has truth value' as it appears in the example in 9.3. So I propose now to add 'proposition has truth value' to clause 8. Since this is a well-known formal logic concept, I doubt that it will be problematic. As a point of interest, the issue about handling 'whether' formulations that are not closed led me to wonder why that was an issue relative to '_closed projection_ means question'. I finally realized that 'closed projection means question' is a fact type that will _never_ be used in a vocabulary, a fact model, or a rule book, or indeed in any body of shared guidance. I suppose it might be used in formulating the meaning of a verbatim quotation. This causes me to wonder whether most of what is in the entry in clause 9.3 would not be better moved to 9.2.9, which deals with questions that will appear in facts and rules. With respect to handling questions involving multiple projection variables, I stand by my position that they need not appear in rules, and the formulation of database queries is out of scope for SBVR. Since it is difficult to create the proper semantics for interpreting the results of such projections as instances of anything defined in SBVR (we don't do table rows), I strongly suggest that we do not try to support them in question formulations. I suppose some additional text will be needed to deal with that. -Ed -- Edward J. Barkmeyer Email: edbark@nist.gov National Institute of Standards & Technology Manufacturing Systems Integration Division 100 Bureau Drive, Stop 8263 Tel: +1 301-975-3528 Gaithersburg, MD 20899-8263 Cel: +1 240-672-5800 "The opinions expressed above do not reflect consensus of NIST, and have not been reviewed by any Government authority." SBVR Issue 10621-d5a.docx