The role of the notion of a proper answer in the logic of questions: Methodological remarks and postulates

The observations in the article mainly concern the role of the concept of the so-called right answer in question logic. The purpose of these remarks is to justify the postulate that any logic of questions should be based on a conception of the structure of questions and answers, in which the notion of a proper answer is strictly defined. This postulate is addressed to any question logic, although it is mainly supported and illustrated by analyses and comparative remarks referring to concepts based on Ajdukiewicz’s question theory and to recent approaches of inferential erotetic logic (IEL). The analyses confirm that the concept of proper answer is fundamental in question theories, as it is assumed in the definitions of almost all concepts relating to questions and answers. In Ajdukiewicz’s concept, it is used explicitly, for example, in the definitions of the conditions of proper questioning and of complete and exhaustive answers. In IEL, it appears explicitly in the definitions of: the pertinent question, the notion of the presupposition of a question (and its variations), the relations of evoking Rocznik Filozoficzny Ignatianum The Ignatianum Philosophical Yearbook Vol. 27, No. 1 (2021), s. 319–340

a question (by a set of indicative sentences) and implying a question (by another question), etc. This basic concept should therefore be well defined. This postulate applies especially to such theories of questions in which assertions about questions and answers are proved in symbolic language -as is the case in IEL, which, however, lacks a strict definition of the concept of proper answer (there are only vague, pragmatic terms formulated in natural language). There is, however, a definition that is closer to the idea of the proper answer, adopted by Ajdukiewicz as well as in the concepts related to it, that a proper answer is one the structure of which is determined by the scheme of the question structure. However, this definition should be complemented by an accurate and general conception of question structure, which is lacking in the existing concepts. In order to confirm the validity of the formulated postulate, the article proposes new results achieved in the theory of questions, in which Ajdukiewicz's ideas are developed and supplemented by a full account of the structure of questions and well-defined, i.e. formulated in a general and strict way as is the idea of proper answer.
Słowa klucze: odpowiedź właściwa, struktura pytań, logika pytań, Ajdukiewicza teoria pytań, inferencyjna logika pytań 1. It can be stated that the concept of a proper answer is present in every logical theory of questions. In any "theory", because this statement does not concern solutions to individual erotetic problems, but conceptions that include: types of questions, conditions for good questioning (especially the accuracy of questions), types of answers and the requirements for answers. And it is "present", as this concept is usually not defined directly, often only used, while in more complex concepts indirectly assumed.
Various terms are used in question theories to denote proper answers, for example, "conclusive answer", "principal possible answer", "sufficient" and "just-sufficient answer", "congruent", "exhaustive", "complete", and "direct answer". 1 This concept (ignoring differences in terminology) is used directly, and it is often also assumed without naming, e.g. when it comes to answers to a given question, when any verbal responses to the given question are distinguished from the expected responses. In informal question theories it is taken explicitly in many definitions; and in formal theories, definitions refer to this concept -usually through symbolic notations used in the definitions of successive, more and more complex concepts. To illustrate the various uses and meanings of this concept, I will point out a few examples taken from Ajdukiewicz's conception of questions and from Inferential Erotetic Logic (IEL). 2

1.1
In concepts referring to Ajdukiewicz, the concept of the proper answer is clearly visible in the postulates on good questioning and when it comes to the kinds of answers and the conditions answers should meet. In this approach and the terminology employed we are talking about questions posed properly/improperly. The necessary and sufficient condition for proper questioning is the truthfulness of its so-called positive assumptions (PA) and negative assumptions (NA): PA states that among the proper answers to a given question there is (at least one) a true answer, and NA -that there is a false answer among them. 3 In turn, when it comes to answer distinctions and postulates, the concept of a proper answer appears when: (i) complete answers (called "full" [całkowite]) are divided by Ajdukiewicz into direct and indirect answers: direct complete answers are equated with proper answers, and indirect complete answers are defined as not being proper, but implying the proper answer to the question; and when (ii) so-called exhaustive answers are defined as true sentences from which each true proper answer follows. 4 1.2 Compared to Ajdukiewicz's conception of questions, IEL is a much more complex and structured theory. Questions and answers formulated in natural language and the regularities observed in the practice of asking questions and answering are reconstructed in a deliberately constructed, symbolic language, the concepts and theorems of this theory being related in terms of definition and proofs. In IEL the concept of a proper answer is most often deemed "direct answer", with the terms "principal possible answer" and "just-sufficient answer" also employed. These terms (one of them) are already visible in relatively simple concepts (definitions), such as the concept of question soundness and question presupposition. By resigning from terms formulated exclusively in the symbolic language of IEL, it can be stated that a question Q is sound if, 2 Since the purpose of this analysis is neither a historical one nor that of furnishing a review of existing literature, references will be limited to Ajdukiewicz's original conception (albeit that it has been employed by many Polish semioticians) and to recent presentations of IFL in [16], [17] and [18], and only if at least one direct answer to Q is true; 5 and the presupposition of the question Q is this and only the sentence that follows logically from each direct answer to Q. 6 The concept of a proper answer ("direct") is also assumed in defining the so-called prospective presupposition of the question Q: it is such an presupposition of the question, the truth of which is a necessary and sufficient condition for the soundness of the question Q. 7 The concept of a proper answer is also a component of more complex concepts, namely the evocation of a question and the erotetic implication. Still resigning from the definitions written in IEL, it can be stated that a question is evoked by a set X of declarative sentences if and only if the truthfulness of each set sentence guarantees that the question Q is sound, but from the set X no specific proper answer logically follows. In other words -if and only if the truthfulness of the sentences of the set X guarantees that the question has a true proper answer, but does not indicate (does not imply) any specific proper answer. 8 As for the erotetic implication, the question Q' is implied by the question Q in the context of the set X of declarative sentences if and only if the truthfulness of the sentences of the set X and the soundness of Q ensure that the question Q' is also sound, i.e. has a true proper answer; and that each proper answer to the implied question Q' narrows the set of proper answers to the implying question (on the basis of the set of sentences X), i.e. it indicates such a proper subset of proper answers to question Q, in which there is at least one true proper answer. 9 The concept of a proper answer is also a component of the definitions of specific kinds of erotetic implications, such as regular, strong and pure erotetic implications. 10 In the symbolic language used in IEL, the concept of a proper answer is referred to by dQ, denoting the set of proper answers to the question Q. 11 The symbol dQ appears directly not only in the strict (in formal language) definitions of the above-mentioned concepts, but in almost every directly formulated (and numbered) definition accepted in IEL, except for the definitions of auxiliary concepts. Therefore, it is reasonable to expect that this notion is defined in a strict manner and in the same language as this entire structure.
2.1 However, there is no such definition. Proper ("direct") answers to the question Q are defined in IEL as those of the possible answers which: " […] provide neither less nor more information than it is requested by Q. Being true is not a prerequisite for being a direct answer. " 12 Answers called "direct" in [18] are in [17] and [16] referred to as Principal Possible Answers (PPA): is a possible answer that is "optimal" in the sense that it provides information of the required kind and, at the same time, provides neither more nor less information than it is requested by the question"; 13 […] direct answers/ppa's are supposed to be the possible just-sufficient answers, where "just-suffient" means "satisfies the request of a question by providing neither less nor more information than it is requested". 14 The definitions of the proper answer (variously termed) proposed in other theories of questions are similar to the above cited. In [4], apart from the "informative" characteristics adopted in IEL, there is a requirement to refer to the question in a direct and precise manner; in [8] it is postulated "an answer which would satisfy the questioner if it were true and if he were in a position to trust the answer. By a conclusive answer, I mean a reply which does not require further backing to satisfy the questioner. "; in [13], proper answers are understood as sentences that anyone who understands a question should consider as acceptable answers to a given question, and at the same time "the simplest, most natural ... "; and according to [7] the "direct" answer provides exactly what the question demands, and directness means both "logical sufficiency and immediacy". These definitions -quoted in IEL to explain and support 10  the adopted characteristics of proper answers 15 -also do not go beyond the "intuitive" level. The notion of a proper answer is defined in them by pragmatic concepts: understanding a question, satisfaction with the answer, trust in the respondent, recognition of the answer as admissible and natural, no need for further justification, the obligation to consider a sentence as an answer to a given question, etc -are pragmatic ones. The requirement for any given answer to provide exactly ("neither less nor more") the required information, if it has not been made precise, also becomes a pragmatic condition; the conditions of logical sufficiency, directness, accuracy and maximal simplicity ("the simplest") of answers also require semantic or syntactic precision. 16 The need for precision is already visible on the "intuitive" level. For example, to possible answers to the question (Q) Who discovered America? undoubtedly belong:

was discovered by Columbus (iv) America was discovered by Christopher Columbus (v) America was discovered by the son of Domenico Colombo, a Genoese weaver and merchant, born in 1451
There are doubts, however, as to which of these and many other possible answers satisfy the condition of providing exactly as much information as is required by the question (Q). For example, the answers (i) -(v) are syntactically different, but all indicate the same person. So if the condition "neither less nor more information" were to be met for question (Q) always and only when exactly one person is indicated among all the possible explorers, then each of the answers (i) -(iv) is an element of d(Q). The same is true for answers (vi) and (vii) for as direct answers these may not be true. However, comparing these answers in the context of the requirement that an answer should not provide too much information, certainly results in the elimination of the answer (viii) from d(Q) determined in this way; and may also lead to the assessment that this requirement is not met in sentence (iv), indicating Columbus through information about his father -and it is easy to formulate even more complex descriptions, yet ones unequivocally signifying Columbus. In turn, the recognition that for this reason (v) ∉ d(Q) casts a shadow of doubt on the answers (iii) and (iv): do they not provide too much, since they contain information known from the question that it is about the discoverer of America? This information is not (directly) in the answers (i) and (ii), but maybe the information about the first name of the discoverer makes the answer (ii) "redundant"? If so, then only the statement (i) or its full sentence equivalent (iii) is an element of the set of "direct" answers, which -according to the assumption adopted in IEL -is uniquely determined by questions (Q). In turn, the rejection of the answer (v), and the inclusion in d(Q) of the answers (i), and (ii) -or (iii) and (iv), (vi) and (vii) -suggests that the condition "neither less nor more" must be understood to mean that it is permissible to use different, but equal-range individual names (Columbus and Christopher Columbus), but it is not permissible to use a general name in the answer, even if denoting the same. 17 2.2 In conceptions based on Ajdukiewicz, the concept of a proper answer is better defined. Apart from pragmatic definitions, a method of characterizing answers that refers to the schemas of questions is employed. Proper answers are defined as sentences that are obtained when, in the answer pattern set by the question (known as datum quaestionis), the variable contained is replaced by a constant from the scope of the variable, i.e. from the range of the so-called unknown of the question. 18 However, such definitions are satisfactory only if it is accompanied by well-defined schemas for any questions. In [3] only the scheme for so-called questions to be completed "consisting of a question particle and a fragment of a declarative sentence" are described in more detail, while for the questions to be completed "[...] which have the entire sentence under the question particle" there is only one example, for the questions to be decided it is only a general hint on building answer patterns, and for questions to be explained, the pattern is omitted. 19 The lack of question 17  schemas is also visible in Ajdukiewicz's analyzes of properly asked questions. Namely Ajdukiewicz indicates how the assumptions PA and NAor better to say the postulates as to the appropriate questioning -should be specified for particular types of questions and on this basis to decide whether a given question is properly asked. Namely, PA is fulfilled if and only if the alternative of the proper answers is true, and NA -if the alternative of the negations of the proper answers is true. 20 However, a study of an appropriate questioning in accordance with these ideas is possible only if the concept of the proper answer is well defined, which in turn requires arrangements as to the structure of the questions and answers of any kind. 21 3. In [12], questions are distinguished from interrogative sentences with possible sources of ambiguity specific to interrogatives exhaustively covered. 22 Sources of ambiguity specific for interrogative sentences -i.e. independent of ambiguity, which may affect any expressions of a given language, including interrogatives -are: (i) the indeterminacy of what is questioned and what is given (this type of ambiguity applies to interrogative sentences for Wh-and Why-questions) (ii) the interrogative sentence quantification, i.e. a requirement, refined in a given asking situation, as to the number of objects from the so-called universe of questions that must be indicated in the answer (concerns W-questions, i.e. Wh-and Why-questions); (iii) the possibility of causal or purposive interpretation of interrogative Why-sentences. 23 3.1 The general scheme of any question structure is given by the formula: (*) ? x* in U*: C*(x*).
In this scheme, "x* in U*" indicates the subject of the question, that is, its unknown and universe, C* is the condition predicated about objects from the universe of the question; and the generally described proper answer to the questions has the form C* (x*), and it must be that x* in In [1] is proposed a general approach to the structure of Wh-questions (and answers to Whquestions), namely the formula ? x P(x) (I owe this information to Prof. A. Brożek). 20 See [3, pp. 88-89]. 21 This gap in Ajdukiewicz's conception has been filled in [14], [13], [15], [9], [18], [6] and [12], among others. 22 Questions are interpreted interrogative sentences, i.e. sentences taken in one of their possible meanings and this distinction is indicated by writing interrogative sentences in plain print, and writing questions based on them in italics (as used in these analyzes). The ambiguity of interrogative sentences is also analyzed in [10]. 23 These interpretations are indicated by the subscripts "c" or "p", respectively: Whycquestione, Whyp-questions.
U*. The symbol in is a variable for which, in less general schemes, one can substitute the symbol of the relationship appropriate for the unknown x* and the universe U*, i.e. ∈ or ⊂ (equality included). For specific questions or kinds of questions, the notation "x* in U*" is detailed according to the subject of the question /kind of question, and the possible quantification of the question can be taken into account. The question condition C* is also properly concretized -up to the concretization appropriate for a given question -and it is possible to take into account the ambiguity of the interrogative sentence, i.e. indicate with the schema exactly this question that expresses one of the meanings of the interrogative sentence being uttered, the appropriate one for a given situation. 24 3.2 For example, the interrogative sentence: 1) Who studies philosophy? if addressed to a specific group of P people, it can be ambiguous only because of ambiguity as to the quantification, i.e. how many people need to be indicated. The scheme for (1) that does not take into account any quantification is the formula: ? x ∈ P: C(x) and for questions with the quantification of n = 1, n = 2, n = k and the quantification of "all", the following schemas are appropriate: ?
In each of these schemas there is the same universe P of people whose names are substituted for the variable x, and the same condition C = studies philosophy visible in the scheme of proper answers (after the colon).
There are, however, Wh-questions whose universe is not the same as the scope of the unknown. This is the case in questions with more than one interrogative pronoun.
2) Who is studying what? If there is a specific set P of people and a set D of disciplines of study that come into play, then also this interrogative sentence may be ambiguous only due to undefined quantification. On the other hand, the subject of questions is different than for (1), visible in the above formulas obtained from (2), because in the scope of the unknown of these questions there are ordered pairs taken from the universe (P × D). Here are the schemas for questions based on (2) -unquantified and quantified in the same way as (1) In these formulas, strings of the type <x, y>k oraz C(<x, y>1, <x, y>2, …, <x, y>k) are abbreviations for <xk, yk> and C(<x, y>1) ∧ C(<x, y>2), ∧, …, C(<x, y>k), the condition C = is studying, and the form of the proper answers is visible after the colon.
In a similar way, it is possible to construct a schema for Wh-questions based on interrogative sentences with more than two pronounsunquantified or with quantifications such as n = ... or "all". On the other hand, questions that are quantified differently, e. g. "at least one", "more than three", "2 ≤ n ≤ 5", etc., must be approached differently. In the question schema, one need to add a condition specifying the number ‖A‖ of the elements of set A in accordance with the quantification required in the question. For example, the shape of the question obtained from (2) for the quantification "not less than k, not more than n" is: ?
It is also possible to grasp which component of the declarative sentence p is concerned by a particular W-question "derived" from p -which in the context of question situations will be called the question matrix p. The corresponding schema indicates exactly what is the universe U, and thus -what is the condition C assumed in a given question, adjudicated in the proper answer about objects from the universe U. Having adopted appropriate notational convention, one can express this difference by the following: ?
Expressions (1a) -(1c) are also written in a way that indicates they are interrogative sentences, because, just like (1), they can give rise to variously quantified questions, which in turn can be given by specifying the number ‖A‖ in the above schemas, and thus in the proper answer schemas. The subscripts indicate that the universe, and hence the condition of the question, is changing. The universes are the specified sets: 3.3 What is common to Why-questions and Wh-questions of the form ? Ap ⊂ Up: Ap = … is that the question segment refers to the whole matrix, i.e. to the sentence p 26 , and that the same quantification can appear in them as in W-questions. However, what distinguishes them is, first, the ambiguity of the question segment Why itself. Such questions require an explanation of what is announced in the sentence p, but because Why can be understood as a question about cause or as a question about purpose, so in the structure schemata for the generally understood why-question -as well as in the proper answer schematathe phrase " ... because ... " appears, and in the versions for the causal and purposive interpretations, respectively: "the reason that ... was that ... " and "the purpose that . is also supposed to indicate that the requirements of a certain grammar -irrelevant in the general analysis -are omitted. (In Polish, it is syntactically correct to place the pronoun Why [Dlaczego] before the sentence p and before any of its constituents). In the reconstruction of Y/N-questions the sign [Y/N] will be used for the same purposes. 29 The combinations assembled from the three components of this matrix are eight; in question situations, seven may appear, because the 0-element combination, i.e. the situation when all the components of the matrix are given (none are questioned), corresponds to the statement p. 30 Writing (3a)-(3e) as questions (indicated by italics) is the result of omitting the quantification that can accompany these expressions; without this simplifying assumption, (3a)-(3e) should be written as question sentences.
The condition C, corresponding to the unquestioned components, is for questions (3b)-(3e) denoted in a way that indicates this, i.e. as C{3}' = Peter studies, C{2}'= Peter philosophy, C{1}'= studies philosophy, C{1, 3}'= studies. Using these denotations, the sentence p can be represented as an ascription: 3}). Each one stands for the sentence p, but in these notations it is clear what the explanation is supposed to be about, and what is unquestioned in the question.
Consistent with the comments and agreements above, the schemata for questions (3b)-(3e), and within them the schemata for proper answers, look thus: 3}) because (x). In turn, the question scheme obtained from (3) when interpreting this sentence according to Why p? is: (3a)' ? x ∈ U{1,2, 3}: C{1, 2, 3}'({1, 2, 3}) because (x). And since {1, 2, 3} = p, this scheme can be simplified: (3a) ? x ∈ Up: p because (x). 31 In a way relativizing to the particular questions, their universes are also labeled in these schemata. There are answers explaining: that philosophy is studied by Peter -e.g. the answer Why, so that after graduation he can teach philosophy; that Peter is connected with philosophy by studying (e.g. Because he is no longer satisfied with studying philosophy on his own); that it is Peter who studies philosophy (Because Andrew has decided to study physics); and answering (3e) one has to explain both that Peter … and that … philosophy, e.g. Because he has earned money for his studies and has always been interested in philosophy. The universes of questions (3a)-(3e) are different, but not disjointed. Indeed, it is worth noting that the answers to questions (3b)-(3e), i.e. questions with universes U{3}, U{2}, 2} and U{1, 3} are also the answers to question (3a), the answers to (3b) and (3d) are the answers to (3e) -that is, in general: answers to a question in which a given component of the matrix p is being questioned are also answers to any other question based on that matrix in which the given component is also questioned. The completeness of a particular answer is a separate matter.
The formulas for questions (3a)-(3e) fall under the general scheme of Why-questions structure: ( In the question schemas, the variable in is replaced by the symbol ∈ or ⊂ , the symbol for the set A may also appear, and a condition specifying the number ‖A‖ may be added -depending on whether and how the particular question is quantified, i.e. how many explanations the questioner expects. Schemata for questions simultaneously determine schemes for proper answers.

3.4
The ambiguity, the source of which is the demarcation between what is questioned and what is given (the question condition), becomes even more evident in the case of Y/N-question sentences. Semiotic analyses of Y/N-questions show that the question segment in a question sentence of the type "Is it so that p?" does not always refer to the whole matrix p, which is a source of potential ambiguity not only for the question itself, but also for a negative answer to such a question. 32 A full analysis of Y/N-question sentences must therefore take into account that in questions -that is, question sentences posed in concrete situations -not only are single components of the matrix p questioned, but also their pairs, triplets, etc., up to the questioning of all components, that is, the whole matrix p.
The result of applying the above way of analyzing the matrix and distinguishing possible meanings of a question sentence to Y/N-interrogatives is a schema where the symbol {...} denotes, as in schemas for questions of other types, the part of the matrix p -from its individual components up to the whole sentence p -that is questioned in a given question from among those obtained from the question sentence "Is it so that p?": (*)Y/N ? x ∈ {{…}, non{…}}: C{…}'(x). 33 32 See, for example: [13, pp. 48-51], [6, pp. 143-144], [16, pp. 9-10]. In Polish, the participle "Czy" suffices in the question segment of any Y/N-questions, both when the entire matrix p is being questioned and in questions concerning any of its parts. 33 [13, pp. 49-51) took into account only the possibility of questioning the individual, single components of the matrix p of the question sentence "Is it so that p?" and proposed a scheme for the structure of such questions, which, in the part concerning what I call the object of the question, is consistent with (*)Y/N. However, the "negative" Accordingly, the universe of each Y/N-question is two-element, the affirmative proper answer follows the formula C{...}'({...}) and is always equivalent to sentence p. In contrast, the negative proper answer, with the scheme C{...}'(non{...}) may not be unambiguous, or more precisely, it is rarely unambiguous. This is because behind C{...}'(non{...}) are hidden unambiguous and complete answers, each of which corresponds to one of the possibilities for negating the segment {...}, and from each such unambiguous answer it follows that C{...}'(non{...}). 34 To facilitate the comparison of schemata, the formula (*)Y/N is applied below to the question sentence corresponding to (3), i.e.
(4) Does Peter study philosophy? and to only those of its seven possible interpretations that have counterparts in the above questions and formulas (3a)-(2e): 35 ( and on the basis of this conception to define the proper answer. A good definition of the concept of the proper answer provides a solid basis for defining more complex concepts and justifying claims about questions and answers. It is necessary, for example, when the presuppositions of a question are defined as sentences that follow logically from each proper answer; a so-called direct presupposition as the logically strongest of the presuppositions of a question; an accurate question as having a true and false proper answer, and so on. On a well-defined concept of the proper answer it is also possible to firmly ground analyses concerning the conditions of accuracy and the relationship between the accuracy of a question and the logical value of its presuppositions, as well as distinctions concerning the kinds of answers and the relationship between them. 39 On the other hand, the initial adoption of the reverse order at the starting point: proper answer -question -question sentence, accompanied by a defective (unclear, vague) definition of the proper answer results in this defect being transferred to the theory. This order of analysis is in IEL 40 , in which questions are then represented by sets of proper answers: ? {A1, …, An}.
In this formula (the so-called e-formula) A1, ..., An are the declarative sentences (which are syntactically distinct in pairs) that are the proper answers ("PPA") to the question. 41 When the reconstruction of questions is based on the notion of a proper answer, one not made precise in the language in which the logical theory of questions is built, then all the definitions and theorems in which this notion is presupposed, for example, the notions basic in IEL, already indicated above, of: the presupposition of a question (and the particularizations of this notion), the soundness of a question (and related notions), the notion of evocation and the variants of erothetic implication, as well as theorems concerning these relations, are deprived of a strict basis. Thus "knowing a question" often yields a disambiguation. " 42 Only if the assumption H2 were true for any question could it justify representing questions by answers. It is certainly not the case that a single answer points unambiguously to a question. Let us suppose, that the sentence counted as an answer is "Socrates discovered America". This sentence is undoubtedly one of the proper answers to Who discovered America? -but not only. It is also included in the set of proper answers to other questions, for example: Socrates discovered America or was he an eminent philosopher? Who among -Socrates, Aristotle, Columbus, Magelan -discovered America?, "What is Socrates known for?", and can also be considered as an abbreviation of the proper answer to "Why is Socrates a famous historical figure?" and as an expansion of the proper answer in the affirmative to "Did Socrates discover America?" 43 Premise H 2 is defensible only if by "knowing what counts as an answer" is meant the necessary and sufficient conditions for any declarative sentence to be decidable as the proper answer to a given question. This, however, requires knowledge of the answer schema (as determined by the question structure schema), which determines the set of proper answers dQ. 44 That assumption H 21 is not true is demonstrated clearly by the fact that knowledge of NLQ's does not lead unambiguously to the formula representing them and that there is no consensus on this in question theories. Assumption H 22 , on the other hand, can be acknowledged, but again: as long as there is an elaborate scheme for the structure of any questions in a given conception. In IEL there is no reconstruction of the structure of questions, because questions are represented (in e-formulas) by sets of answers; in the language of this theory, therefore, "knowing a question" has not been achieved.

4.3
A postulate that logic of questions should be based on a concept of question structure (taking into account the ambiguity of question sentences) but is, however, confronted with the fact that not only has no theory of questions been agreed upon, but also no structure conception itself. This fact fosters doubts as to whether an accurate and general question structure conception is possible at all; and -how to prove its accuracy and generality. These requirements and related doubts apply, of course, also to the question structure conception outlined above. For example, how can it be verified that any question, i.e. any of the possible meanings of any question sentence formulated in a given natural language (not to mention any natural language), can be subsumed into the question patterns according to (*), i.e., (*)U, (*)W and (*)Y/N and their further specifications. Proving that for any given natural language question it is the case that it falls under one of these schemes, and thus also under (*), is not possible. On the other hand, empirical verificationi.e. corroborating the results of a logical theory of questions with facts drawn from the practice of posing and answering questions -encounters otherwise known limitations. Confirming the thesis of the generality of these schemata involves effectively pressing further, specific questions into them. A better way than random questions testing is to test such questions, which previous erothetic analyses have pointed out as difficult to reconstruct and classify, i.e. to show that the anomalies of other question theories disappear. It is also valuable to show that other categorizations can be reduced to the one obtained on the basis of the proposed schemes that "inter-theoretical reduction" is possible. Indeed, selected other classifications can be reduced to schemes based on (*), which confirms that the division of the universe of questions into two categories, i.e. Y/N-questions and others (W-questions), among which there are Wh-questions and Why-questions, originating from K. Ajdukiewicz, is correct. 45 4.4 Related to the above is another methodological postulate: the logic of questions should firstly be pragmatic, and then formal. This postulate allows one to stop -as with Ajdukiewicz -at pragmatics and formulate the concept of questions in natural language enriched with a few methodological terms; however, it is supposed to protect against developing a formal logic of questions without confronting it with the empirical base drawn from the practice of asking and answering. The assumptions made in the formal logic of questions -both at the starting point and for proving the theorems intended therein -must not be incompatible with this practice. Only then can the results obtained in the formal logic of questions in turn provide a basis for explaining, normalizing and improving this practice.