Discovering Jan Smedslund's Psychologic: Challenging the Assumptions of Psychology Forfatter:

The current paper starts with a presentation of the metatheoretical landscape of current psychology. Jan Smedslund's psychologic is presented as a post-modern, or constructionalist, approach to psychology as a science. The key elements of psychologic are presented and explained. The system is then applied to two randomly chosen psychological articles: 1) a study of the predictive value of phonemic awareness in kindergarten children for later reading success; 2) an investigation of how autonomous motives influence physical activity intentions within the Theory of Planned Behaviour. Both of these articles were found to benefit from the application of psychologic. Further, the critics of Smedslund are surveyed, and those that pertain to this paper's analyses are examined in detail. It is concluded that the system of psychologic provides conceptual clarity and logical structure. The price for this conceptual clarity is the loss of flexibility that accompanies the rigid definitions in psychologic. It is argued that the application of psychologic can serve as a useful tool for explicating hypotheses and uncovering necessary relationships in empirical studies, but that it does not constitute a substitute for empirical research. ___________________________________________________________________________ Author: Stud. Psychol. Fredrik Pedersen. Institute of Psychology, University of Tromsø. Title: Discovering Jan Smeslund's Psychologic: Challenging the Basic Assumptions of Psychology Subject: Psychology. Grade: Master thesis for the degree of Cand. Psychol. November, 2004. Discovering Psychologic 4 Discovering Jan Smedslund's Psychologic: Challenging the Basic Assumptions of Psychology "Is psychology properly viewed as being 100% empirical? ... If there really are no limits to what we might discover in the course of doing "scientific psychology", then there would also be no limits to how we might properly describe what we discovered and no limits to the concepts we might use in explaining or understanding what we discovered. But this would hold true if we set out to do 100% empirical physics or 100% empirical economics ... or 100% empirical anything. So then, why in the world would we call it psychology?" (Shotter, 1991, p. 352) The quote from Shotter points out that in psychology, as in all science, empirical investigations spring from presuppositions that are taken for granted, and help us understand what we are studying. Jan Smedslund (1997) has tried to explicitly state what we must take for granted when dealing with psychological phenomena. The resulting axioms are formulated in his psychologic (Smedslund, 1997). Armed with his psychologic, Smedslund claims to show that a lot of prominent empirical psychological research merely demonstrates what we must take to be necessarily true, given the way we talk about psychological phenomena (Smedslund, 1978; 1984; 1990; 1991a; 1999a; 2002). In this paper, Smedslund's alternative method of inquiry, psychologic (Smedslund, 1997), will be outlined and scrutinized. After an introduction to the metatheoretical landscape the theory grew out of, a summary of the theory will be presented. Two randomly selected articles will then be analysed using psychologic, in order to separate conceptually necessary hypotheses, from empirical hypotheses. Finally, the possibilities and limitations inherent in psychologic are discussed in light of the criticisms the theory has attracted. Tracing the Common Ground in Psychology There are many areas of investigation that fall under the "psychology" umbrella, from exploring at interrelationships between people to examining the interrelationships between neurotransmitters. Despite this diversity, there are some prototypical psychological undertakings. One of the defining features of psychology since it began its professional history in the laboratories of Wilhelm Wundt (Alexander, 2003; Hergenhan, 1997) has been the application of a methodological perspective borrowed from the natural sciences. The goals of the scientific discipline of psychology can be found explicitly defined in introductory textbooks in methodology: to describe, predict, determine the causes of, and understand human behaviour (Cozby, 1993; Shotter, 1991). These goals implicitly embrace Discovering Psychologic 5 several meta-theoretical preconceptions regarding humans as subjects of study. Objectivity (i.e. the world as objectively given, independent of the perceptions and interpretations of individuals) and universality (i.e. cross-context and cross-time validity of observed causeeffect relationships) are two of the key assumptions of science. These epistemological beliefs go hand in hand with an empirical approach to explaining human behaviour. Given that our observations are accurate reflections of the world around us, it makes sense to observe behaviour in conditions that minimize possible errors in interpretation in order to uncover universal cause-effect relationships. The prototypical manifestation of this type of investigation is the experiment. The experimental way of collecting data, with its focus on non-participation and objectivity, has from the beginning separated the academic discipline of psychology from everyday psychology, as well as from its philosophical predecessors. This method of investigation has remained characteristic for the field to date (Danziger, 1990). In an ideal experiment, all observed dissimilarity can be attributed to the manipulation of the variable in question. However, these ideal conditions are rare, even in the physical sciences, and under obtainable conditions there are usually several sources of variance. Of specific importance to psychology is the between-subject variation that is inherent in nearly all human responses. In psychology, statistical analysis represents a coherent logic for making decisions regarding the numerical characteristics of a population when one is in possession of a sample of scores. For example, in the analysis of variance test (ANOVA), the between group variance is divided by the within group variance. This is done in order to see whether groups of subjects are more different from one another than subjects within one group. If groups are found to be significantly different from one-another, one concludes that some of the variance can be attributed to the samples being drawn from different populations. Statistical analysis in psychology can be complicated, and some researchers (e.g. Danziger, 1990; Gigerenzer, 1998) have voiced the concern that the pronounced focus on methodology has gone too far, and that questions of methodological stringency has come at the expense of other worthwhile investigations. One such concern is the extensive focus on quantity in psychological research at the expense of the quality of the psychological phenomena under investigation. According to Danziger; "Concern with questions of methodological orthodoxy often takes the place of concern about theoretical orthodoxy when research or its results are discussed and evaluated." (1990, p. 5). Discovering Psychologic 6 "Modern" Empiricism versus "Post-modern" Constructionalism The assumptions underlying an experimental approach to gathering information can be categorized as belonging to a modern epistemological perspective. During the last half of the twentieth century, a growing body of critics has voiced the opinion that science in general, and psychology in particular, may be labouring under false premises in that the modern perspective, despite its claims to the contrary, does not yield accurate, neutral descriptions of the world. Kuhn (1970) was one of the first to describe some of the social pressures that apply to the "objective" world of science. In his book "The Structure of Scientific Revolutions" Kuhn (1970) describes how a scientific community itself plays a central role in shaping the way phenomena are understood. According to Kuhn, to learn to participate in a scientific community is to study the past scientific achievements that are considered the foundation for further practice. In doing this, students are introduced to, and learn to do research within, the current scientific paradigm. Paradigms proscribe to a certain way of understanding different phenomenon. Shared commitment to one paradigm ensures that its practitioners engage in the observations that this paradigm can do most to explain (Kuhn, 1970). Later writers have focused more directly on the way in which our pre-understandings of what we observe are constitutive of the phenomena we end up describing (Gergen & Davis, 1985). This post-modern epistemological perspective challenges the foundations of modernity, asserting that the world is as much in the eye of the beholder as it is "out there"; that science, rather than being a rational, objective way of gathering knowledge, is a social endeavour. What one "discovers" has to adhere to the rules of the specific scientific community one belongs to with regard to the investigatory practices utilized, and the nomenclature employed to describe the observation (Gergen, 2001). In addition, the object of study must be consensually evaluated as worthy of attention. Psychologists embracing this post-modern epistemological perspective refer to themselves as constructionalists, emphasising the constitutional role of pre-understanding in constructing whatever phenomenon one deals with (Gergen, 1985). The Constitutional Role of Language The role of language is central to the constructionalist perspective and its assail on the epistemology of modernism. Within science (including psychology), language is often used unreflectively (Smedslund, 1997). This means that while phenomena are communicated and understood in terms of language, language itself is not in focus (Smedslund, 1997). The social constructionalist perspective proposes a radical change in the way the role of language is Discovering Psychologic 7 understood (Gergen, 2001). According to this perspective, language is the primary means by which we learn to navigate the social world. In acquiring language, the individual takes part in a social activity and a cultural practice that transcends the individual's limited history. By learning how to articulate and describe different phenomena within a culture one is simultaneously introduced to the "shared reality" of that culture. To be able to meaningfully express oneself means to take part in the shared intelligibility of a culture, including the nomenclature associated with different phenomena, as well as adhering to beliefs regarding their interrelationships. In this sense, the words of a language are not merely "labels" attached to real world phenomena; they play a central role in determining how we come to interpret and understand what we see (Gergen, 2001). From the constructionist perspective, an analysis of our shared conceptual framework (i.e. language) must precede any attempt at description of human behaviour. If our concepts play a constitutive role in shaping how we see and understand the world, then an analysis of these concepts could inform us of phenomena just as much as observation could. According to Smedslund (1991; 1997; 1999; 2002) this insight is generally overlooked by psychologists, and much research is therefore misguided. Lack of Conceptual Definitions as the Source of "Pseudoempiricism" Smedslund, along with other researchers, has questioned whether a substantial amount of empirical psychological research is, in fact, investigations that only appear to contribute to psychological knowledge but in reality are of a tautological nature, not contributing with anything that could not be known without the research (Smedslund, 2002; Wallach & Wallach, 1998). Rigidly constructed experiments and precise statistical analysis can be a wasted effort when the concepts under investigation are muddled with inaccuracy, circularity, or when the "independent" and "dependent" variables are, in fact, interdependent by virtue of their conceptual associations (Smedslund, 2002). Shotter (1991) argues that psychological research has an abundance of examples in which an insufficiently defined concept inspires research leading to confusion rather than clarification, resulting in appeals for more rigorous theoretical analysis to bring order in the chaos. Smedslund agrees, and states that due to lack of conceptual analysis, psychologists often conduct experiments in which they assume that the outcome must be empirically established when, in fact, the hypothesis in question must be regarded as true, independently of empirical demonstrations. Smedslund argues that if the concepts under study are interdependent, the appropriateness of the hypothesis is given beforehand, or a priori. Discovering Psychologic 8 To illustrate this point, Smedslund (2002) invites us to take part in a hypothetical experiment investigating whether people who are surprised have just experienced something unexpected. According to Smedslund, surprise may well be defined as "the state of a person who has just experienced something unexpected" (Smedslund, 2002, p. 52). This conceptual relation means that the hypothesis under investigation is true by virtue of logical necessity given that the words mean what they mean. If our experiment fails to find this relation then one of the auxiliary hypotheses connected to our procedure (e.g. that the instrument used will be appropriate to detect and measure expectancy and surprise) has been disproved, but not the main hypothesis (that surprised people have just experienced something unexpected). Thus, experiments constructed to test necessarily true propositions really only test the accuracy of the methods involved. The main hypothesis in this type of experiment is true by virtue of conceptual necessity, and cannot be strengthened or weakened by empirical investigations. Smedslund (1984; 2002) goes on to argue that in order for there to exist an empirical relation between variables that is not given beforehand, these variables need to be conceptually unrelated. If this is so, then any possible combination of outcomes is both possible and plausible. This, in turn, makes real hypothesis testing possible, where both the main-, and auxiliary hypothesis are tested. Smedslund argues that there are very few areas of investigation in psychology that warrant empirical investigation. Further, the true empirical propositions that can be found within psychology have little but local validity (Smedslund, 1984). The reasons for this are that all psychological phenomena are historical and that historical processes always contain a random component and, hence, are irreversible. Individuals are to a substantial amount a product of their histories. The histories of individuals are punctuated with arbitrary events, and as a result of this, each individual becomes unique and hence incomprehensible and unexplainable except by reference to a series of unique historical events. If one aims to create theories of human action, there has to be some regularity upon which one can build a theory (Backe-Hansen & Schanke, 2004). Smedslund contends that one such source of regularity can be found within language (Smedslund, 1991a; 1997; 1999a; 2002). Smedslund's Psychologic The feature of language that Smedslund (1997) directs attention to is its inherent structure. Although any statement can be uttered, there are clear limitations to what can be Discovering Psychologic 9 meaningfully said, given that the words mean what native speakers of a language take them to mean. If spoken by a fellow human being, the sentence "I am not a person" is hard to comprehend. If the speaker is not a person, then who is the "I" that he/she is referring to? At the very least, this statement informs the listener that some vital context that might render the proclamation understandable is missing. Common Knowledge as the Starting Point for Conceptual Analysis Common knowledge was disregarded as a source of knowledge in the early stages in the professional history of psychology. A probable reason for this was that psychology needed to establish itself as an autonomous project, independent from its philosophical predecessors. Fritz Heider (1958) was one of the first psychologists to see that whereas the relationship between common knowledge and science is generally viewed as one where the latter is superior in its command of the truth, this relationship may be more balanced within the field of psychology. In Heider's own words: "Psychology holds a unique position among the sciences. "Intuitive" knowledge may be remarkably penetrating and can go a long way toward the understanding of human behavior [sic], whereas in the physical sciences such common-sense knowledge is relatively primitive. If we erased all knowledge of scientific physics from our world, not only would we not have cars and television sets and atom bombs, we might even find that the ordinary person was unable to cope with the fundamental mechanical problems of pulleys and levers. On the other hand, if we removed all knowledge of scientific psychology from our world, problems in interpersonal relations might easily be coped with and solved much as before." (Heider, 1957, p. 2) In keeping with Heider's respect for common knowledge, Smedslund (1991; 1997; 1999; 2002) argues that psychology must look to the intuitive knowledge embedded in our use of language. As mentioned, within language there are limits as to what can be meaningfully said. There must be some system in language that makes it apparent when statements doand when they do notcohere. Another way of stating the previous is that language consists of concepts that are interrelated, and can be combined in meaningful ways. Every statement implies some other statements, and negates yet others. For example, being a person implies having a physical body; being a student implies that one attends some form of study, and so forth. Competent users of a language presumably agree about these implications. This shared intelligibility of implications is what Smedslund refers to as common sense (Smedslund, 1984; 1999). Negations of these common sense principles are contradictory or senseless given that the words mean what they mean. It is important to emphasise that this differs from another popular definition, where common sense is viewed as Discovering Psychologic 10 predictions and explanations provided by lay people, and, hence can be subject to empirical investigation. According to this latter definition, scientific language is an improvement over the vernacular. Part of the common sense of a language (i.e. shared intelligibility of implications) refers to psychological phenomena. This sets the conditions for what can be meaningfully said about psychology. Smedslund's psychologic (1997) represents an attempt to explicitly state the structure of this psychological common sense that constitutes the social reality in which people live. Smedslund (1991; 1997; 2002) argues that language, in order to function to coordinate social activity, has to be understood in the same way by a large number of people. Words are not only defined by their context, but many (if not all) words have a core meaning that will be understood by native speakers of the language (Smedslund, 1991; 1997; 2002). Even if a word is removed from the context of other words and presented to a person, it will not be devoid of meaning. The word "surprise" will inspire similar definitions for native speakers; it has to do with the experience of something that was not expected. According to Smedslund, language thus represents a common ground from which one can make valid generalizations about native speakers of a language. Logic fundamental to understanding In outlining the underlying structure of common sense, it became clear to Smedslund that to a significant extent, the inherent structure in common sense is describable using terminology borrowed from logic. The reason for this is that logic is one attempt at explication and formalization of the limits of what can be meaningfully and coherently uttered. As we shall see, Smedslund argues that all understanding presupposes logic (1990). In elaborating this argument, one must start by answering the following questions: What does it mean to understand something? How can one explicate understanding? Smedslund (1991b) explains that these very questions were the ones that inspired the formulation of what was to become one of the first axioms (see below) of psychologic. In an interpersonal context one can say that: "[person] P understands what [another person] O means by saying or doing [something] A, if, and only if, P and O agree as to what, for O, is equivalent to A, implied by A, contradicted by A, and irrelevant to A (Smedslund, 1997, brackets added). It should be emphasised that this statement applies to understanding of what someone means. There are two important implications of this. The first one is that understanding, the way it is formulated here, is a matter of degree, and can never be complete. Discovering Psychologic 11 The second one is that through this statement, one can comprehend what Smedslund (1990) refers to as the circular relationship between understanding and logic. An elaboration of this argument follows. When inspecting whether a person has understood something correctly, the only way to check his/her understanding is by inferring this from their judgements of equivalence, implication, negation and irrelevance. In this inference, one must assume that the individual has used proper logic when reaching his/her conclusion. On the other hand, one can only decide whether a person has used logic correctly by tracing the premises to the conclusion, thus taking understanding of the premises for granted (for an extensive elaboration of this argument, see Smedslund, 1990). According to Smedslund (1990), this interrelationship entails that the properly illogical can never be explained, nor understood. "Explaining involves describing premises from which a given something follows, and understanding involves describing what follows from a given something" (Smedslund, 1990 p. 116). Since the properly illogical does not follow from any premises, explanation and understanding are impossible. Smedslund argues that what we call fallacies (i.e. errors in reasoning), are subjective misunderstandings of the premises that merely appear to be logical errors when one compares the subjective performance to some objective standard. Smedslund argues that when one investigates the subjective premises (i.e. the premises as the individual has understood them) further, in order to ascertain why the individual did not comply with expectations of optimal performance, it invariably turns out the premises are misunderstood. Smedslund (1990) argues that if the aim of psychology is to understand and explain human behaviour we have no option but to presuppose that all voluntary action follows logically from the premises of the individual, and thus, to understand why a person acts the way he or she does, we need to understand these individual premises. The Structure of Psychologic Even if one agrees that closer inspection of our psychological concepts is a necessary step toward understanding human behaviour, this endeavour is not without its problems. An analysis of language needs to be communicated by language to reach others. Thus, this analysis would be embedded in, and limited by, the very concepts it attempts to explicate. This could be seen as analogous to trying to inflate a balloon from the inside: without some external input, endless circularity seems unavoidable. Psychologic has introduced primitive terms in order to stop this circularity. Discovering Psychologic 12 Psychologic in its current form consists of 22 primitive terms, 43 definitions and 55 axioms. It is proposed as an axiomatic system that consists of basic principles which are assumed to be true. The tenets of psychologic are claimed to be consensually self evident. This entails that they express truths that necessarily follow from the shared meanings of the terms involved. It is proposed that all native speakers of the language regard these tenets as self-evidently true, and their negations as senseless. Primitive terms are terms that are not further defined, but are considered basic and self-evident in that they cannot be meaningfully or better defined using other terms. Primitive terms establish a core in the psychologic system that the theory can develop from. Smedslund (1997) argues that, ultimately, all logical systems need a core that cannot be further reduced. Without this core, any attempt to define something would lead to reduction "ad absurdum", or circular definitions (Smedslund, 1997; Wierzbicka, 1992). The selection of primitive terms in psychologic has in part been based on Wierzbicka's work on a natural semantic metalanguage (Smedslund, 1997). To get a feel for the structure of this system, I will give a short presentation of a primitive term, a definition, and an axiom. A mere listing of the terms of psychologic does not do justice to the thoroughness with which terms are treated in the system. As such, I have chosen to give the examples in their entirety. One of the first primitive terms introduced in psychologic is act (do) (Smedslund, 1997, p. 4). Along with the introduction of this term, Smedslund notes: "In encountering a person we take it for granted that the person acts, does things, or expresses him or herself, in order to reach goals, looks and listens in order to determine what is going on, and so on. The person is continuously sensitive to the outcome of these activities, which means adjusting subsequent activities in the light of what resulted from earlier ones. This characteristic of acting is labelled intentionality." (Smedslund, 1997, pp. 4-5) Definitions in psychologic are limited to introducing technical/scientific terms that the reader may not be familiar with. Definition 1.2.3 states that intentional = directed by a preference for achieving a goal (Smedslund, 1997). Axioms are postulates that are taken for granted by any native speaker of the language. As in any axiomatic system, the axioms in psychologic are assumed to be independent, that is they cannot be derived from one another, and consistent, that is they do not lead to contradiction. There has been a move away from definitions to axioms in the development of psychologic (Smedslund, 1997). The reason for this is that to try to capture the meaning of a word in a classical definition is to specify the entire meaning of the word across contexts. Discovering Psychologic 13 When it comes to words from ordinary language, they have a richness that cannot be captured in terms of a classical definition. Axioms do a better job of freezing the content of a concept only in relation to other concepts, not per se. Axiom 1.2.4 states that acting is intentional (Smedslund, 1997). In addition to the above, psychologic contains corollaries that are deduced from propositions involving primitive terms, axioms and definitions. From the preceding we get corollary 1.2.5: Acting is directed by a preference for achieving a goal, and corollary 1.2.6: If [a person] P does something not directed by a preference for achieving a goal, then what P does is not acting (Smedslund, 1997). Along with the above corollaries, Smedslund (1997) notes that this use of "intentionality" differs from usage in normal language. In normal language something is said to be done intentionally if it is preceded by an intention or decision to act. In psychologic, intentional activity refers to all activity that involves preference for achieving a goal, and hence, all activity that is sensitive to outcomes. Thus we get the somewhat counterintuitive result that coughing is seen as an act if it ceases when reproached, but not an act if it continues unaffected by this outcome. If a person's hand is shaking and this is not affected by outcome (e.g. spilling coffee), then the shaking is not an act and must be attributed to something outside the persons control or awareness (such as Parkinson's disease). An Application of Psychologic Although Smedslund has shown his theory to be applicable to several empirically established psychological phenomena (Smedslund, 1978a; 1984; 1991a; 1999a; 2002), I wanted to evaluate for myself whether the hypothesis generated by mainstream psychologists of today are merely "pseudoempirical" investigations into necessarily true propositions. The following demonstration cannot prove or disprove the tenets of psychologic, but it may serve as a potent demonstration of the usefulness of psychologic. Selection For this demonstration, two articles were randomly chosen from the PsycInfo. Eligible articles had to conform to the following criteria: a) published in or after the year 2000; b) be concerned with a psychological issue; c) obtainable in Norway (i.e. with BIBSYS reference); and d) written in Norwegian or English. Articles were chosen by searching the PsycInfo database (through WebSPIRS 5.0) for accession numbers drawn from a random number table (Cozby, 1993, p. 283). The search was entered as "200*-xxxxx-00x" where "*" is a wildcard, Discovering Psychologic 14 and "x's" are digits from the random number table. For example, from the random number 13 74 67 00 78, the six first digits were entered into the formula as: "200*-13746-007". In this case an article was hit (Mason et al., 2004), but it was discarded as it was from the British Medical Journal, and as such failed to satisfy criterion b. A second article was discarded for failing to satisfy criterion d because it was written in German. Study 1 Random number 09 37 67 07 15, hit the article by Wilde, Goerss, and Wesler (2003), entitled "Are all phonemic awareness tests created equally? A comparison of the Yopp-Singer and the task of auditory analysis skills (TAAS)". This study is a comparison between two tests of phonemic awareness, and their ability to predict later reading success (Wilde et al., 2003). Phonemic awareness is defined as "one's sensitivity to, or explicit knowledge of, the phonological structure of words in one's language. In short, it involves the ability to think about, or manipulate the individual sounds in words" (Wilde et al., 2003, p. 295). Phonemic awareness has been shown to be a strong predictor of reading success, and it has been proposed that children should be screened for phonemic awareness in order to identify children who might benefit from extra interventions when learning to read. The two tests of phonemic awareness (Test of Auditory Analysis Skills-TAAS, and Yopp-Singer Test of Phoneme Segmentation) were administered to 25 second-semester kindergarten children. The dependant variable was the STAR-reading test, a computer-based reading test that provides reading scores for students in Grades 1-12. The STAR was administered to the children in February of their First-Grade year. The Yopp-Singer test asks children to listen while the experimenter pronounces a word. The child is then asked to reproduce each sound sequentially. So, if the word is "dog", the correct reply would be /d/ /o/ /g/. When scoring, no partial credit is given so the child's score is determined by how many words were correctly segmented. The TAAS asks children to repeat a word (e.g. "cowboy"), and then to repeat the word but to omit one of the syllables (e.g. "cow" or "boy"). Alternatively, the target word is one syllable (e.g. "meat"), and the child must omit a phoneme from the word (e.g. if asked to omit /m/ a correct reply would be "eat", or if asked to omit /t/ a correct reply would be "me"). The study found that the Yopp-Singer test did not correlate significantly with the STAR-reading test. The TAAS was positively correlated with STAR (r=.51). The children's results on the Yopp-Singer and the TAAS were also positively correlated (r=.56) Discovering Psychologic 15 Study 1, analysed using psychologic Psychologic is a system of interand intrapersonal common sense. As such the processes of learning to read and write are not specifically dealt with. The following analysis will therefore be based on applying the principle of the a priori, rather than draw on applying actual psychologic propositions. This means that the study will be analysed to see whether the hypothesis in question is, in fact, an empirical hypothesis. To qualify as empirical the concepts involved in the hypothesis cannot be related by definition or conceptual relation (Smedslund, 1984; 1991a; 1999a; 2002). The scope of the study by Wilde et al. (2003) was to see whether the Yopp-Singer and the TAAS both predict reading success equally well. The hypothesis that phoneme analysis will predict reading is (with reference to empirical studies) taken as a given. However, the fact that phoneme analysis is correlated with reading can also be known a priori, as will be argued below. The process of phoneme analysis, defined as the ability to identify the phonemes of a word, is an integral part of writing. This is a necessity given the way our written language is based on the auditory properties of spoken language. This is not the case for logographic systems of writing, or in older pictographic writing, and this argument does not apply to these languages (Fisher, 2001). One can also take as a given that phoneme synthesis, defined as the ability to identify separate sounds as a word (e.g. "combine the sounds /d/ /o/ /g/ into one word"), will be correlated with reading success. This is referred to as phonics by Wilde et al. (2003). However, since the synthetic and the analytic processes are interwoven (one can hardly imagine knowing one without knowing the other), one can assume that any instrument measuring either one will be correlated with reading success. If a person is not able to separate the phonemes of a word, the person will be unable to learn to read and write our written language by reference to its depiction of sounds. It should be mentioned that the preceding argument does not apply to learning to read through word recognition. The idea that phoneme analysis/synthesis is predictive of reading success presupposes that phoneme analysis is learned prior to reading and writing. To some extent, this is necessarily true as one cannot understand the principle of our written language without first, or at least simultaneously, being introduced to the concept of words consisting of many separate sounds. However, there seems to me to be no way to state a priori whether this ability exists independent of learning to read and write. Thus, this matter of deciding whether or not phonemic analysis is learned prior to writing, and hence can be used to predict writing success, represents a true empirical hypothesis. Discovering Psychologic 16 The study by Wilde et al. unfortunately does not offer an answer to the predictive qualities of phonemic awareness, as children were not tested for reading skills simultaneously with tests of phonemic awareness. The test of reading skills (STAR) was administered separately long after the TAAS and the Yopp-singer. On account of this, when looking at this study in isolation, the study cannot speak to the issue of prediction. Phonemic awareness, as measured by the TAAS, was correlated with later reading success. Phonemic awareness, as measured by the Yopp-Singer was not correlated with later reading success. Although the present correlation cannot speak to the predictive value of the TAAS, this relationship needs further empirical investigation. As for the Yopp-Singer, the present study eliminates it as a possible tool for early prediction of reading success. In summary; the application of psychologic to the study by Wilde et al. (2003) shows that the question of whether phonemic awareness predicts reading success is conceptually sound. It has been argued that phonemic awareness is an integral part of learning to read, given the way our written language is constructed, the issue of whether this skill is established prior to learning to read must be established empirically. Further, the validation of specific procedures and instruments will always be an empirical endeavour that cannot be anticipated by conceptual analysis. However, one can argue that such validation would profit from a design where the main hypothesis is given a priori so that the result can be unambiguously attributed to the appropriateness of the auxiliary hypothesis (Howard, 1991; Smedslund, 2002). Study 2 Random number 15 95 33 47 64, hit the article by Hagger, Chatzisarantis, and Biddle (2002), entitled "The influence of autonomous and controlling motives on physical activity intentions within the Theory of Planned Behaviour". This study investigated how general goals (i.e. autonomous and controlling motives) influence physical activity intentions within the framework of the Theory of Planned Behaviour (Hagger et al., 2002). The Theory of Planned Behaviour (TPB) is a model of human reasoning that has been applied to questions of health related behaviour. The model presupposes that humans are rational decision makers and that our intention to engage in any behaviour is based on considerations of our attitudes toward the behaviour in question, our subjective norms, and our perceived behavioural control (PBC). Hagger et al. (2002, p. 284) explain that in TPB "intention is viewed as the most proximal predictor of behaviour ... and is hypothesized to mediate completely the influence of the cognitions of attitude, subjective norms and PBC on Discovering Psychologic 17 behaviour". Attitudes in TPB are aggregates consisting of the belief that the behaviour in question will produce a given outcome, and the value placed on this outcome. Subjective norms consist of the individual's beliefs regarding whether or not other people want him/her to perform a given behaviour, and the individual's motivation to comply with these referents. Finally, PBC refers to the individual's belief that he or she is capable of performing the behaviour in question. The Theory of Planned Behaviour is summarized in Figure 1. Figure 1. The Theory of Planned Behaviour. Adapted from G. Smedslund (2000a, p. 136) Hagger et al. (2002) argue that there is a weakness in TPB in that it does not account for the influence of superordinate, cognitive influences on behaviour. Although the TPB helps to identify what cognitive factors underlie intentions to act, it may be limited in that it is context-specific. Hagger et al. suggest that the TPB would be improved by adding the influence of more general motives on the cognitive factors of the TPB. They argue that constructs from self-determination theory can be adopted to demonstrate that "individuals can base their intentions ... on the higher, more general motives generated by their psychological need for self-determination" (p. 285). The person's belief that the behaviour leads to certain outcomes and his/her evaluation of these outcomes The person's belief that specific individuals or groups think he/she should or should not perform the behaviour, and his/her motivation to comply with specific referents. Attitude toward the behaviour