A problem statement---any science statement whose truth value is being questioned (Hars, 2003, p.169), because of disconfirming facts or because it is incoherent with inferred statements. This is formalized here by the requirement that statements have a 'problem-in' facts that are ‘disconfirmed-by’ other facts, or are incoherent with other inferred statements. A science problem signals the need for a revision of a theory or fact, such that the disconfirmations or incoherencies are resolved. Note that the notion of disconfirmation and incoherence also takes into account the inconsistency between theories and models: (1) theory: the science role that is being disconfirmed is a theory part; (2) model: the fact responsible for the disconfirmation can represent data about an observed particular that can exist in models; or the prediction responsible for the inconsistency can refer to a hypothetical particular that can exist in models. 1 The functional role played by a theory or its part. e.g. some theory might assume a statement (Assumption), others might target the statement as a problem (ScienceProblem), and others might not believe it to be only locally applicable (Empirical Regularity), while in others its considered a science law and universally applicable. A statement can play different roles in different theories, and can evolve to play different roles in time within a single theory. For example, a statement that is an empirical regularity could evolve to a law with more verification, and any statement could play the role of a science problem if disconfirmed via facts or shown to be logically inconsistent with some theory part. Also, science roles cannot be discovered independent of being produced, because the products of science are newly discovered knowledge elements; hence 'discovered-by' is not allowable for science roles, while its inverse of 'discovery' is allowed because it refers to the particulars that are discovered as a result of the scientific statement. Science roles are related in two ways to a theory: (1) as a role played by a whole theory (or some other description) or by its part, or (2) as a component of a whole theory. The first involves role-playing while the second involves a whole-part relation. Thus a science role can be played-by by certain theories or its parts, or it can be part of certain theories. 0 1 A science theory that is satisfied only by a geoscience model. An observation or measurement activity that produces observed data. Implicit in this is the idea that only qualities of physical-endurants or perdurants can be observed, but this is relaxed such that the physical-endurants or perdurants can also be observed. 1 The definition of a concept in some description, such as in a science theory. May include reference to other concepts (e.g. to define necessary or sufficient condition) and/or prototypical instances. A collection of particulars (instances) that are satisfied by a science theory, where each particular is discovered by a Fact or Prediction or Proposition, and is predictable (can be deduced) from a Prediction that uses the science theory that is satisfied by the model. The science model itself is produced by a scientific modeling activity. In a logical sense satisfaction implies the presence of some interpretation (or denotation) function between a syntax + grammar and a model, but in a scientific sense satisfaction implies here that a particular in a model can be deduced from a theoretic role (proposition, regularity, law) that is part of the satisfying theory. For example, in the case of classification a particular passes a classification rule (theoretic role) associated with a concept, which places the particular in the extension (model) for the concept. The classification rule can be based on similarity to examples, necc. & suff. conditions, or some other criteria. In SKIo, the rule can be associated with the concept as a concept definition that plays the role of proposition, regularity or law within a theory. For a quick survey see 'Models in Science': http://plato.stanford.edu/entries/models-science/. There, models in science have 2 roles which can overlap: (a) representing some aspect of reality, or its data, or (b) instantiating (satisfying) a theory. Lee (1999) also looks at types of models. The spirit of (b) is adopted here (as per DOLCE): a model is a collection of instances, which satisfy one or more theories. This is also totally consistent with (a) since a theory in most cases only predicts some aspects of reality and cannot make statements about others. It is also consistent with (a) because in DOLCE a model is a representation of reality. Some implications: - a 'mathematical model'--a set of a related equations--is a theory. - a mathematical equation is a statement within a theory (a part of a mathematical theory). - the set of data which satisfies the mathematical theory is a model - a physical model of the world (e.g. made with Lego) is a model which satisfies some theory (i.e. an architectural philosophy or movement (art deco?)) - in a computer simulation of physical reality: the configuration of physical entities constitute the model, the code, types and rules that govern entity behavior are its theory. - the geospatial objects on a map are a model, and the legend consisting of class definitions is a theory (Brodaric & Hastings, 2002) Some similar views: Sowa (2000): a model in Logic consists of a collection of entities in a 'world' as well as a denotation function that maps those entities to terms in logical statements such that the statements evaluate to TRUE--the collection of statements form a theory. The model is extensional, representing the world, and the theory is intensional, representing abstractions about the world. (a model instantiates a theory) Shrager and Langley (1990): a model is required to indicate the manner in which a law or theory applies to a particular situation. (a model instantiates a theory) 1 A reasoning step performed by some rational agent as part of some plan. For our purposes this includes both sentient biological agents and machine agents. This assumes all inferences follow a plan, in the sense that they follow some prior strategy or pattern, rather than in the sense that they are predicted or predetermined. For example, a plan might be some rendition of the general scientific method in which the different inference types are used within specific stages--e.g. abduction is used for theory discovery, deduction for predictions about the world, and induction for prediction testing (Sowa, 2000). There 3 basic types of inference, according to Sowa (http://www.jfsowa.com/talks/souprepr.htm): 1. Induction or learning: Start with observations. Look for commonalities. Derive a theory that summarizes the data. 2. Deduction or inference: Start with a theory. Observe some new data. Use the theory to deduce the implications. 3. Abduction or guessing: Start with disconnected observations. Guess (hypothesize) a theory that relates them. Test the hypothesis by induction and deduction. 1 Involves generating a Proposition to enable coherence of discordant scientific knowledge elements. SKIo Abduction encompasses both logical and pragmatic abduction. Logical abduction is reverse deduction: in a deduction of the form T ^ a1 ├ b1, where T and b1 are known, then guess the missing instance a1. Pragmatic abduction, on the other hand, is more concerned with guessing theory T via conceptual transformations from radically different prior knowledge (often via analogy ~): e.g. given T1:CD, and A~C ^ B~D, then T2:AB. SKIo Abduction encompasses both logical and pragmatic abduction, such that the resultant Proposition refers to either a1 (an instance) or T2 (a theoretical relation). 1 A basic geoscientific object. Refers to any solid, oceanic or atmospheric physical body that is produced as a consequence of natural processes, rather than social or mental processes. It is disjoint with biological and chemical physical bodies. Examples include rock bodies, bodies of water, bodies of air, etc. A description typically containing the qualities associated with some particulars. Data about physical and temporal qualities can often be observed, while data about abstract qualities (associated with e.g. social objects) are typically inferred (from observed data). Although, strictly speaking, only qualities are observable, for generality data includes descriptions of any particular, including endurants, perdurants, qualities and abstracts. Data is more specific than a Description only in that Data describes some specific particular as a result of some activity leading to the description. 1 The product of an observation, which is typically some description of the observed values of some physical or temporal quality. It is internally represented by some observing agent, e.g. the scientist doing fieldwork, or the memory chip of a remote sensing unit. More specifically, observed data is the representation in some description of the values of physical qualities inherent in physical endurants or the value of temporal qualities inherent in perdurants. For flexibility, observed data can also be seen to describe the physical endurant or predurant that exhibit those qualities. 1 2 1 Involves generating a prediction about the world using existing theory and data. Logically, given theory T:AB (A and B are concepts) and instance a1 (of A), then b1 (of B) is deduced: T ^ a1 ├ b1. In SKIo Deduction, T is realized as a theoretical science role (Proposition, EmpicalRegularity, ScienceLaw), a1 as a science role about an observed or inferred instance (Fact, Prediction, Proposition), and b1 as the resultant Prediction. Pragmatic Deduction: a prediction about the state of the world made from some existing theory and data. Pragmatic and Logical deduction are completely aligned, hence Deduction here encompasses both. The product of Deduction is a prediction, which is a statement about the state of the world, generally (in the geosciences) about the values of qualities of instances of physical endurants or perdurants An activity that produces a scientific assumption. The assertion may be performed on some existing scientific statement to indicate the statement is being treated as TRUE. 1 1 0 1 Finding patterns in data (logical induction), or dis/confirming a pattern via data collection (pragmatic induction). Here, in SKIo, induction is explicitly used in the first sense as a means of finding patterns in data, and implicitly used in the second sense via the verification activity which dis/confirms a scientific statement (relative to a fact) or finds that it in/coheres with other statements. In logical induction: given data { (a1, b1), (a2, b2), (a3, b3) } then infer T (A, B), where T is some theoretical relation over concepts A, B, and ai and bi are their respective instances. In pragmatic induction: given T (A, B), note in newly observed data { (a1, b1), (a2, b2), (a3, b3) } and infer T (A, B) ╞ TRUE (or FALSE if disconfirmed), or given a1 note in data { a1, a1, …} and infer a1 = TRUE (or FALSE if disconfirmed). SKIo Induction refers to logical induction performed on Data where the induced T is an EmpiricalRegularity or ScienceLaw, while Verification encompasses pragmatic induction resulting in dis/confirmation of a science role. In addition to testing against data, Verification also encompasses testing for in/coherency of theoretical relations against each other. Intuitively, a non-process-like geoscience entity: an endurant consisting of the union of geoscience physical bodies, amounts of matter that constitute such bodies, and features that host geoscience bodies. Used to constrain the contents of geoscience models to these entities. The physical expression (encoding) of some method, which can be executed by a computing device. 1 Scientific statements express a scientific idea as a relation amongst scientific knowledge elements (after Hars, 2003, p 144), in some real scientific artefact such as a paper, report, map, etc. 1 1 1 1 1 1 1 The entity in which statements are published. Publication refers loosely to any entity that is accessible and that hosts science statements. Contains only a representative suite of properties, as this is just a placeholder for a more extensive metadata schema--this is also why the properties are all strings rather than Objects (e.g. authored-by could have as range Agent rather than string). 1 A perdurant with only GeoScienceEndurants as participants, i.e. a natural process or event. 1 A science role that is produced via an inference instead of asserted or observed. The discovery relation refers to a particular that is discovered as a result of this inference, e.g. a specific rock body or process that is discovered as a result of the inference. The activity of testing scientific statements against Facts, or against other theoretic statements (InferredRoles). This activity could result in added properties to a tested statement, indicating the statement is dis/confirmed-by or rendered in/coherent-by the verification activity. Note that testing against Facts can implicitly equate to testing against models, inasmuch as the particular being described by the data playing the Fact can exist in models. Of the four types of verification listed below, only prediction testing is explicitly represented in SKIo: (1) Testing Fact vs Fact: involves checking whether observations are consistent. Implies one or both Facts can be dis/confirmed by the other. (2) Testing Fact vs a role about individuals (Predictions, Propositions): involves Prediction testing, such that the Prediction can be dis/confirmed by the fact, or Proposition testing such that the Proposition can be disconfirmed by the fact. Note for Prediction testing this further implies the deduction components are possibly dis/confirmed by the Fact--i.e. the theoretical entity (Proposition, Regularity, ScienceLaw) and empirical entity (Fact, Prediction) from which the prediction is deduced are thus in question and one or both could be dis/confirmed by the Fact. (3) Testing Fact vs theoretic role (Proposition, ScienceRegularity, or ScienceLaw): involves checking whether a Prediction can be deduced from the theory such that the Prediction is dis/confirmed by the Fact, i.e. whether the Fact falls into the deductive closure of the theoretic role (i.e. theorem proving). (4) Testing theoretic roles against each other: involves doing consistency checking for logical consistency. Another form of testing here is model-checking, where theoretic roles are consistent (coherent) if they satisfy the same models, which is similar to (3) in the sense that scientific satisfaction entails that the model component can be deduced from the theoretic role, but where the model component is not necessarily observed (i.e. is either observed (Fact) or inferred (Prediction, Proposition)). A testable conjecture about observable entities (perdurants and endurants, and their qualities), that is the product of a deduction (Sowa, 2000). It can be empirically tested and must be logically consistent, or coherent, with some existing theory or theories (after Hars, 2003), which are usually used to originally deduce the prediction. 1 The general activity of producing science models, which are collections of particulars that satisfy science theories. Could include 'simulation' as a subtype. Because science models are collections of particulars in which each particular is discovered-by a fact or prediction, science modeling is the activity of configuring these particulars into a whole such that the whole satisfies some science theory. In a logical sense satisfaction implies the presence of some interpretation (or denotation) function between a syntax + grammar and a model, but in a scientific sense satisfaction implies here that a particular in a model can be deduced from a theoretic role (proposition, regularity, law) that is part of the satisfying theory. For example, in the case of classification a particular passes a classification rule that places the particular in the extension for a concept. The classification rule can be based on similarity to examples, necc. and suff. conditions, or some other criteria. In SKIo, the rule can be represented as a concept definition that plays the role of proposition, regularity or law within a theory, and the extension can be represented as a science model. 1 Facts are roles played by data as part of some science theory. In this case facts are always "theory-laden" in the sense that they are intrinsically part of some theory. In terms of basic characterization, facts are an empirical statement supported by observed data, i.e. a statement about the real world that has been verified scientifically via observation. They are often considered to refer to a single endurant or perdurant (Hars 2003), but this restriction is not formalized here. The discovery relation in a fact refers to a particular that is discovered as a result of this fact, e.g. the value of a quality of an endurant (rock, water body, etc.) or the endurant that is itself identified in the observation. 1 The activity of identifying a scientific problem. Scientific problems are identified when certain scientific statements are designated as being problematic--these are statements which are disconfirmed-by empirical facts or are found to be incoherent with other relevant theory. Generally speaking such disconfirmations and inconsistencies are produced by verification. 1 A geoscience model is a collection of geoscience particulars that satisfy a geoscience theory. Geoscience particulars include geoscience physical bodies, amounts of matter, features that host geoscience bodies, and perdurants in which geoscience bodies participate. An empirical pattern that is highly confirmed by many observations and that is universal. It relates concepts whose instances are observable. (after Hars, 2003; Shrager and Langley, 1990). The pattern is often stated as a relation amongst concepts. This relation holds globally for objects at a certain granularity, such that the relation is always present when the instances of its concepts are present. This universality is formalized here via negation: the existence of a science law is not dependent on any specific observable entities. 1 The activity of taking samples. The activity of carrying out analysis of samples or data, using an instrument and possibly some software, resulting in data being produced. The physical expression (encoding) of a collection of data. The same data could be encoded using different physical expressions--i.e. there could be syntactic or structural differences (but not semantic differences). 1 AKA 'entity'.Any individual in the DOLCE domain of discourse. The extensional coverage of DOLCE is as large as possible, since it ranges on 'possibilia', i.e all possible individuals that can be postulated by means of DOLCE axioms. Possibilia include physical objects, substances, processes, qualities, conceptual regions, non-physical objects, collections and even arbitrary sums of objects.The class 'particular' features a covering partition that includes: endurant, perdurant, quality, and abstract. There are also some subclasses defined as unions of subclasses of 'particular' for special purposes: spatio-temporal-particular (any particular except abstracts)- physical-realization (any realization of an information object, defined in the ExtendedDnS ontology). A particular is an abstraction about single entities (see Masolo et al, 2003). There are 4 main subtypes in DOLCE: abstracts, qualities, endurants and perdurants. Particulars can be prototype instances for some concept definition, and all prototypes (except science roles) can be discovered by scientific means, through inference or observation. Science roles are not discoverable, rather they are producible via scientific activity. Production (product-of) and discovery (discovered-by) are distinct here: production refers to activity by which some particular is created, whereas discovery refers to the activity by which some particular is realized to exist. For scientific roles, the two are equivalent, i.e. product-of = discovery-by, because the products of science are newly discovered knowledge elements. Hence, discovery is disallowed from science roles. Also, when a particular is a member of science model it is deducible from the science theory that is satisfied by the model. A statement that is considered to be a primitive within a theory, i.e. it is declared via an assertion without empirical proof and without being necessarily inferred from some other element. To be maximally useful, assumptions should be accompanied by roles that describe the function of the statement within the theory, e.g. if a science law is assumed to be true it is likely advantageous to include the original science law role (which is inferred), to enhance overall explanation of the statement. 1 1 An empirical pattern that is produced by induction, but which is valid only for a subset of the possible entities that could satisfy it; i.e. it is a correlation that holds locally amongst some empirical entities, but does not necessarily hold universally for all the empirical entities to which the regularity might apply. If the regularity is expressed as a relation amongst concepts, then the relation is present only amongst some situations in which instances of the concepts are jointly present, and not in all situations where the concept instances are jointly present. This is formalized by the notion that the existence of the regularity is dependent on specific instances, but not all valid instances. It might not hold universally because of insufficient confirmation (i.e. confirmed only in a few situations), or because the pattern can only occur in a local situation (specific space-time region) because it is an artifact of a particular historical process. In general, a theory is a collection of coherent (broadly self-consistent) scientific statements that describe the structure or behavior of some aspect of reality in sufficient generality such as to be applied in a wide number of real-world instances or models (after Hars, 2003). See notes on 'ScienceModel' for more about models and theories. Specifically, a science theory, or part of a theory, is here an entity satisfied only by a science model. A science theory can then refer to a whole theory (the container) or to any of its parts. The functional role the part plays in the theory is specified by its ScienceRole. A theory (container) also keeps track of all of the science roles played by itself and its parts, via the component relation. The scientific discovery of the whole theory or a part is usually specified in the role, via the product-of relation, but could equivalently be specified by the discovered-by relation. It is possible then for the theory to play a science role, which can specify how the role and thus theory is produced, and for the theory to be discovered-by the same science role—in both cases the net effect is to specify how the theory comes to be. Refers here to a specific research activity, or a collection of activities, such as undertaken during a research project. Research produces descriptions (such as theories or data), information objects (such as databases or publications), or science models. 1 A conjecture generated from an inference by abduction (Sowa, 2000), which is not necessarily empirically verified, and from which verifiable hypotheses can be generated (Hars, 2003). Propositions may contain any concepts. Propositions may not be empirically proven, but should be logically consistent, or coherent, with existing relevant theories. Propositions might be valid locally or universally; i.e. they may or may not be situated, that is, dependent on specific endurants or perdurants. Scientific propositions are analogous to logical propositions in the following way: scientific propositions are eventually verifiable (are testable against the world) while logical propositions are eventually resolvable to True or False (are testable against a logical system). A physical instrument which is used to make a scientific observation or measurement. Involves testing a prediction against a fact (data in the context of some theory). The result is a confirmation or disconfirmation of the prediction. This is achieved by adding a confirmed-by or disconfirmed-by relation to the prediction, one that has as range the relevant verification activity. 2 A relation between an activitiy and an some agent (or agency) supplying funding for the activity. A relation between a sciene role (Fact or InferredRole only) and the verification activity in which the science role is dis/confirmed against facts or found to be in/coherent with other inferred roles. A relation between a science role (Fact or InferredRole only) and the activity in which the science role has been verified to be inconsistent against some facts. A relation between an InferredRole and an activity in which the role has been verified to be consistent with some other InferredRoles. A relation between a science problem and some science roles that exhibit are disconfirmed ot incoherent, usually identified as such as a result of a verification activity. A relation between an InferredRole and an activity in which the role has been verified to be inconsistent with some other InferredRoles. A relation between a science role (Fact or InferredRole only) and the activity in which the science role has been verified to be consistent against some facts. A relation between an activity and the entity upon which the activity is performed (i.e. the receiver of the activity), e.g. in an observation this is the entity being observed or measured. A relation between a science role (Fact or InferredRole only) and a particular that is discovered in relation to the science role, e.g. a rock body is discovered via observation (a Fact) or via inference (Prediction). The particulars that satisfy this prediction. A relation between a particular and some prediction from which particular (or its properties) can be deduced. A relation between the motivation (of an agent) and the activity being performed (by the agent).