Use of the familiarity difference cue in inferential judgments

The familiarity difference cue has been regarded as a general cue for making inferential judgments (Honda, Abe, Matsuks, & Yamagishi in Memory and Cognition, 39(5), 851–863, 2011; Schwikert & Curran in Journal of Experimental Psychology: General, 143(6), 2341–2365, 2014). The current study tests a model of inference based on familiarity differences that encompasses the recognition heuristic (Goldstein & Gigerenzer, 1999, Goldstein & Gigerenzer in Psychological Review, 109(1), 75–90, 2002). In two studies, using a large pool of stimuli, participants rated their familiarity of cities and made choices on a typical city-size task. The data were fitted with the r-s model (Hilbig, Erdfelder, & Pohl in, Journal of Experimental Psychology: Learning Memory and Cognition, 37(4), 827–839, 2011), which was adapted to include familiarity differences. The results indicated that people used the familiarity difference cue because the participants ignored further knowledge in a substantial number of cases when the familiarity difference cue was available. An analysis of reaction-time data further indicated that the response times were shorter for heuristic judgments than for knowledge-only-based judgments. Furthermore, when knowledge was available, the response times were shorter when knowledge was congruent with a heuristic cue than when it was in conflict with it. Differences between the familiarity difference cue and the fluency heuristic (Schooler & Hertwig, 2005, Psychological Review, 112, 610–628) are discussed.     

Picture (im)perfect: Illusions of recognition memory produced by photographs at test

Photographs have been found to affect a variety of psychological judgments. For example, nonprobative but semantically related photographs may increase beliefs in the truth of general knowledge statements (Newman, Garry, Bernstein, Kantner, & Lindsay, Psychonomic Bulletin & Review, 19(5), 969–974, 2012; Newman et al., Journal of Experimental Psychology: Learning, Memory, and Cognition, 41(5), 1337–1348, 2015). Photographs can also create illusions of memory (Cardwell, Henkel, Garry, Newman, & Foster, Memory & Cognition, 44(6), 883–896, 2016; Henkel, Applied Cognitive Psychology, 25(1), 78–86, 2011; Henkel & Carbuto, 2008). A candidate mechanism for these effects is that a photograph increases the fluency with which a statement or an event is processed. The present study was conducted to determine whether photos at test can induce illusions of recognition memory and to test the viability of a conceptual fluency explanation of these effects. The results of the present study suggest that photographs enhance the fluency of related words (Experiment 1), that false memories can be produced by the mere presence of a related photo on a recognition memory test for words (Experiments 2 & 3), and that these effects appear to be limited to conceptually based recognition tests (Experiments 4 & 5). The results support the notion that photograph-based illusions of memory stem from the ability of related photographs to increase the speed and ease of conceptual processing.     

Modus Ponens Under the Restrictor View

There is a renewed debate about modus ponens. Strikingly, the recent counterexamples in Cantwell (Theoria, 74, 331–351 2008), Dreier (2009) and MacFarlane and Kolodny (The Journal of Philosophy, 107, 115–143 2010) are generated by restricted readings of the ‘if’-clause. Moreover, it can be argued on general grounds that the restrictor view of conditionals developed in Kratzer (1986) and Lewis (1975) leads to counterexamples to modus ponens (cp. Charlow Synthese, 190, 2291–2323 2013; Khoo Philosophical Studies, 166, 153–64 2013). This paper provides a careful analysis of modus ponens within the framework of the restrictor view. Despite appearances to the contrary, there is a robust sense in which modus ponens is valid, owing to the fact that conditionals do not only allow for restricted readings but have bare interpretations, too.     

Temporal encoding strategies result in boosts to final free recall performance comparable to spatial ones

The method of loci is a highly effective mnemonic that recruits existing salient memory for spatial locations and uses the information as a scaffold for remembering a list of items (Yates, 1966). One possible account for the effectiveness of the spatial method of loci comes from the perspective that it utilizes evolutionarily preserved mechanisms for spatial navigation within the hippocampus (Maguire et al. in Proceedings of the National Academy of Sciences, 97(8), 4398–4403, 2000; O’Keefe & Nadel, 1978; Rodriguez et al. in Brain Research Bulletin, 57(3), 499–503, 2002). Recently, though, neurons representing temporal information have also been described within the hippocampus (Eichenbaum in Nature Reviews Neuroscience, 15(11), 732–744, 2014; Itskov, Curto, Pastalkova, & Buzsáki in The Journal of Neuroscience, 31(8), 2828–2834, 2011; MacDonald, Lepage, Eden, & Eichenbaum in Neuron, 71(4), 737–749, 2011; Mankin et al. in Proceedings of the National Academy of Sciences, 109(47), 19462–19467, 2012; Meck, Church, & Matell in Behavioral Neuroscience, 127(5), 642, 2013), challenging the primacy of spatial-based functions to hippocampal processing. Given the presence of both spatial and temporal coding mechanisms within the hippocampus, we predicted that primarily temporal encoding strategies might also enhance memory. In two different experiments, we asked participants to learn lists of unrelated nouns using the (spatial) method of loci (i.e., the layout of their home as the organizing feature) or using two novel temporal methods (i.e., autobiographical memories or using the steps to making a sandwich). Participants’ final free recall performance showed comparable boosts to the method of loci for both temporal encoding strategies, with all three scaffolding approaches demonstrating performance well above uninstructed free recall. Our findings suggest that primarily temporal representations can be used effectively to boost memory performance, comparable to spatial methods, with some caveats related to the relative ease with which participants appear to master the spatial versus temporal methods.     

The Hidden Set-Theoretical Paradox of the <Emphasis Type="Italic">Tractatus</Emphasis>

We are familiar with various set-theoretical paradoxes such as Cantor's paradox, Burali-Forti's paradox, Russell's paradox, Russell-Myhill paradox and Kaplan's paradox. In fact, there is another new possible set-theoretical paradox hiding itself in Wittgenstein’s Tractatus (Wittgenstein 1989). From the Tractatus’s Picture theory of language (hereafter LP) we can strictly infer the two contradictory propositions simultaneously: (a) the world and the language are equinumerous; (b) the world and the language are not equinumerous. I call this antinomy the world-language paradox. Based on a rigorous analysis of the Tractatus, with the help of the technical resources of Cantor’s naive set theory (Cantor in Mathematische Annalen, 46, 481–512, 1895, Mathematische Annalen, 49, 207–246, 1897) and Zermelo-Fraenkel set theory with the axiom of choice (hereafter ZFC) (Jech 2006: 3–15; Kunen 1992: xv–xvi; Bagaria 2008: 619–622), I outline the world-language paradox and assess the unique possible solution plan, i.e., the mathematical plan utilizing ‘infinity’. I conclude that Wittgenstein cannot solve the hidden set-theoretical paradox of the Tractatus successfully unless he gives up his finitism.     

Probabilistic Opinion Pooling with Imprecise Probabilities

The question of how the probabilistic opinions of different individuals should be aggregated to form a group opinion is controversial. But one assumption seems to be pretty much common ground: for a group of Bayesians, the representation of group opinion should itself be a unique probability distribution (Madansky [44]; Lehrer and Wagner [34]; McConway Journal of the American Statistical Association, 76(374), 410–414, [45]; Bordley Management Science, 28(10), 1137–1148, [5]; Genest et al. The Annals of Statistics, 487–501, [21]; Genest and Zidek Statistical Science, 114–135, [23]; Mongin Journal of Economic Theory, 66(2), 313–351, [46]; Clemen and Winkler Risk Analysis, 19(2), 187–203, [7]; Dietrich and List [14]; Herzberg Theory and Decision, 1–19, [28]). We argue that this assumption is not always in order. We show how to extend the canonical mathematical framework for pooling to cover pooling with imprecise probabilities (IP) by employing set-valued pooling functions and generalizing common pooling axioms accordingly. As a proof of concept, we then show that one IP construction satisfies a number of central pooling axioms that are not jointly satisfied by any of the standard pooling recipes on pain of triviality. Following Levi (Synthese, 62(1), 3–11, [39]), we also argue that IP models admit of a much better philosophical motivation as a model of rational consensus.     

Recall is not necessary for verbal sequence learning

The question of whether overt recall of to-be-remembered material accelerates learning is important in a wide range of real-world learning settings. In the case of verbal sequence learning, previous research has proposed that recall either is necessary for verbal sequence learning (Cohen & Johansson Journal of Verbal Learning and Verbal Behavior, 6, 139–143, 1967; Cunningham, Healy, & Williams Journal of Experimental Psychology: Learning, Memory, and Cognition, 10, 575–597, 1984), or at least contributes significantly to it (Glass, Krejci, & Goldman Journal of Memory and Language, 28, 189–199, 1989; Oberauer & Meyer Memory, 17, 774–781, 2009). In contrast, here we show that the amount of previous spoken recall does not predict learning and is not necessary for it. We suggest that previous research may have underestimated participants’ learning by using suboptimal performance measures, or by using manual or written recall. However, we show that the amount of spoken recall predicted how much interference from other to-be-remembered sequences would be observed. In fact, spoken recall mediated most of the error learning observed in the task. Our data support the view that the learning of overlapping auditory–verbal sequences is driven by learning the phonological representations and not the articulatory motor responses. However, spoken recall seems to reinforce already learned representations, whether they are correct or incorrect, thus contributing to a participant identifying a specific stimulus as either “learned” or “new” during the presentation phase.     

Pinocchio against the Semantic Hierarchies

The Liar paradox is an obstacle to a theory of truth, but a Liar sentence need not contain a semantic predicate. The Pinocchio paradox, devised by Veronique Eldridge-Smith, was the first published paradox to show this. Pinocchio’s nose grows if, and only if, what Pinocchio is saying is untrue (the Pinocchio principle). What happens if Pinocchio says that his nose is growing? Eldridge-Smith and Eldridge-Smith (Analysis, 70(2): 212-5, 2010) posed the Pinocchio paradox against the Tarskian-Kripkean solutions to the Liar paradox that use language hierarchies. Eldridge-Smith (Analysis, 71(2): 306-8, 2011) also set the Pinocchio paradox against semantic dialetheic solutions to the Liar. Beall (2011) argued the Pinocchio story was just an impossible story. Eldridge-Smith (Analysis, 72(3): 749-752, 2012b) responded that unless the T-schema is a necessary truth of some sort (logical, metaphysical or analytic), the Pinocchio principle is possible. Luna (Mind & Matter 14(1): 77–86, 2016) argues that the Pinocchio contradiction proves the principle is false. D’Agostini & Ficara (2016) discuss a more plausible physical truth-tracking trait, the Blushing Liar, and argue that the Pinocchio contradiction is not a metaphysical dialetheia. I respond to Luna, and D’Agostini & Ficara, and prove that the Pinocchio paradox is a counterexample to hierarchical solutions to the Liar.     

Lost in Translation: Religion in The Public Sphere

This paper proposes a Wittgenstein-inspired critique of the prism of translation that frames the recent literature about the debate between Rawls and Habermas on the role of religious reasons in the public sphere (Habermas 2008; Weithman 2006; Wolterstorff 1997). This debate originates with the introduction of Rawls’s proviso in his conception of the public use of reason (Rawls The University of Chicago Law Review, 64(3), 765-807, 1997), which consists in the “translation” of religious reasons into secular ones, which he thinks is necessary in order for religious reasons to be legitimate in the public sphere (Courtois Dialogue, 49, 91-112, 2010; Loobuyck and Rummens Ars disputandi: The Online Journal for Philosophy of Religion, 5, 237–249, 2011; Sikka The Review of Politics, 78, 91-116, 2016). Even though Wittgenstein is not himself concerned with religious pluralism as a political issue, there are numerous scholars who have discussed the political implications of his remarks (Gunnell Contemporary Political Theory, 12 80-101, 2013; Livingston Philosophy and Social Criticism, 33(6), 691–715, 2007; Moore Philosophy and Social Criticism, 36(9), 1113-1136 2010; Pohlhaus and Wright Political Theory, 30(6), 800–27, 2002). The thesis of this paper is that the interpretation proposed by Cora Diamond (2000) in regards to ethical and religious questions turns out to be a suitable way out of the “translation requirement”. According to this solution, if there is to be an understanding between secular and religious citizens on the basis of religious reasons, it should not rely on a “translation” but rather on mutual self-representation.     

Norms of valence and arousal for 14,031 Spanish words

Most current models of research on emotion recognize valence (how pleasant a stimulus is) and arousal (the level of activation or intensity that a stimulus elicits) as important components in the classification of affective experiences (Barrett, 1998; Kuppens, Tuerlinckx, Russell, & Barrett, 2012). Here we present a set of norms for valence and arousal for a very large set of Spanish words, including items from a variety of frequencies, semantic categories, and parts of speech, including a subset of conjugated verbs. In this regard, we found that there were significant but very small differences between the ratings for conjugations of the same verb, validating the practice of applying the ratings for infinitives to all derived forms of the verb. Our norms show a high degree of reliability and are strongly correlated with those of Redondo, Fraga, Padrón, and Comesaña’s (2007) Spanish version of the influential Affective Norms for English Words (Bradley & Lang, 1999), as well as those from Warriner, Kuperman, and Brysbaert (2013), the largest available set of emotional norms for English words. Additionally, we included measures of word prevalence—that is, the percentage of participants that knew a particular word—for each variable (Keuleers, Stevens, Mandera, & Brysbaert, 2015). Our large set of norms in Spanish not only will facilitate the creation of stimuli and the analysis of texts in that language, but also will be useful for cross-language comparisons and research on emotional aspects of bilingualism. The norms can be downloaded and available as a supplementary materials to this article.     

Estimating the average need of semantic knowledge from distributional semantic models

Continuous bag of words (CBOW) and skip-gram are two recently developed models of lexical semantics (Mikolov, Chen, Corrado, & Dean, Advances in Neural Information Processing Systems, 26, 3111–3119, 2013). Each has been demonstrated to perform markedly better at capturing human judgments about semantic relatedness than competing models (e.g., latent semantic analysis; Landauer & Dumais, Psychological Review, 104(2), 1997 211; hyperspace analogue to language; Lund & Burgess, Behavior Research Methods, Instruments, & Computers, 28(2), 203–208, 1996). The new models were largely developed to address practical problems of meaning representation in natural language processing. Consequently, very little attention has been paid to the psychological implications of the performance of these models. We describe the relationship between the learning algorithms employed by these models and Anderson’s rational theory of memory (J. R. Anderson & Milson, Psychological Review, 96(4), 703, 1989) and argue that CBOW is learning word meanings according to Anderson’s concept of needs probability. We also demonstrate that CBOW can account for nearly all of the variation in lexical access measures typically attributable to word frequency and contextual diversity—two measures that are conceptually related to needs probability. These results suggest two conclusions: One, CBOW is a psychologically plausible model of lexical semantics. Two, word frequency and contextual diversity do not capture learning effects but rather memory retrieval effects.     

The influence of object similarity and orientation on object-based cueing

Responses to targets that appear at a noncued position within the same object (invalid–same) compared to a noncued position at an equidistant different object (invalid–different) tend to be faster and more accurate. These cueing effects have been taken as evidence that visual attention can be object based (Egly, Driver, & Rafal, Journal of Experimental Psychology: General, 123, 161–177, 1994). Recent findings, however, have shown that the object-based cueing effect is influenced by object orientation, suggesting that the cueing effect might be due to a more general facilitation of attentional shifts across the horizontal meridian (Al-Janabi & Greenberg, Attention, Perception, & Psychophysics, 1–17, 2016; Pilz, Roggeveen, Creighton, Bennet, & Sekuler, PLOS ONE, 7, e30693, 2012). The aim of this study was to investigate whether the object-based cueing effect is influenced by object similarity and orientation. According to the object-based attention account, objects that are less similar to each other should elicit stronger object-based cueing effects independent of object orientation, whereas the horizontal meridian theory would not predict any effect of object similarity. We manipulated object similarity by using a color (Exp. 1, Exp. 2A) or shape change (Exp. 2B) to distinguish two rectangles in a variation of the classic two-rectangle paradigm (Egly et al., 1994). We found that the object-based cueing effects were influenced by the orientation of the rectangles and strengthened by object dissimilarity. We suggest that object-based cueing effects are strongly affected by the facilitation of attention along the horizontal meridian, but that they also have an object-based attentional component, which is revealed when the dissimilarity between the presented objects is accentuated.     

Interpreting confidence intervals: A comment on Hoekstra,Morey, Rouder,and Wagenmakers (2014)

Hoekstra, Morey, Rouder, and Wagenmakers (Psychonomic Bulletin & Review 21(5), 1157–1164 2014) reported the results of a questionnaire designed to assess students’ and researchers’ understanding of confidence intervals (CIs). They interpreted their results as evidence that these groups “have no reliable knowledge about the correct interpretation of CIs” (Hoekstra et al. Psychonomic Bulletin & Review 21(5), 1157–1164 2014, p. 1161). We argue that their data do not substantiate this conclusion and that their report includes misleading suggestions about the correct interpretations of confidence intervals.     

Ontological Issues and the Possible Development of Cultural Psychology

Ontological issues have a bad reputation within mainstream psychology. This paper, however, is an attempt to argue that ontological reflection may play an important role in the development of cultural psychology. A cross-reading of two recent papers on the subject (Mammen & Mironenko, Integrative Psychological and Behavioral Science, 49(4), 681–713, 2015; Simão Integrative Psychological and Behavioral Science, 50, 568–585, 2016), aimed at characterizing their respective approaches to ontological issues, sets the stage for a presentation of Cornelius Castoriadis’ ontological reflections. On this basis, a dialogue is initiated with E.E. Boesch’s Symbolic Activity Theory that could contribute to a more refined understanding of human psychological functioning in its full complexity.     

Imperfection,Indecision, and Hoarding

Hoarding disorder is a new DSM-5 disorder that causes functional impairment and affects 2 to 6% of the population (Frost and Steketee 2014). The current study evaluated a multiple mediation model with 243 undergraduate women in which indecisiveness (VOCI; Thordarson et al. Behaviour Research and Therapy, 42(11), 1289-1314, 2004) and decisional procrastination (DPS; Mann 1982) mediated the relationship between dimensions of perfectionism (F-MPS-B; Burgess et al. 2016a) and hoarding behavior (SI-R; Frost et al. Behaviour Research And Therapy, 42(10), 1163–1182, 2004) and excessive acquiring (CAS; Frost et al. Annual Review of Clinical Psychology, 8, 219–242, 2012). Multiple mediational analyses indicated a significant indirect effect for decisional procrastination, but not indecisiveness, in mediating evaluative concerns (but not striving) to SI-R Total, SI-R Clutter, SIR Excessive Acquisition, and both CAS subscales. Both mediators were significant pathways between evaluative concerns and SI-R Difficulty Discarding. These findings support a cognitive behavioral model of hoarding, suggesting that evaluative concerns produces problems in decision-making that influence acquisition, discarding, and clutter.     

TISK 1.0: An easy-to-use Python implementation of the time-invariant string kernel model of spoken word recognition

This article describes a new Python distribution of TISK, the time-invariant string kernel model of spoken word recognition (Hannagan et al. in Frontiers in Psychology, 4, 563, 2013). TISK is an interactive-activation model similar to the TRACE model (McClelland & Elman in Cognitive Psychology, 18, 1–86, 1986), but TISK replaces most of TRACE’s reduplicated, time-specific nodes with theoretically motivated time-invariant, open-diphone nodes. We discuss the utility of computational models as theory development tools, the relative merits of TISK as compared to other models, and the ways in which researchers might use this implementation to guide their own research and theory development. We describe a TISK model that includes features that facilitate in-line graphing of simulation results, integration with standard Python data formats, and graph and data export. The distribution can be downloaded from https://github.com/maglab-uconn/TISK1.0.     

Modulation of additive and interactive effects by trial history revisited

Masson and Kliegl (Journal of Experimental Psychology: Learning, Memory, and Cognition, 39, 898–914, 2013) reported evidence that the nature of the target stimulus on the previous trial of a lexical decision task modulates the effects of independent variables on the current trial, including additive versus interactive effects of word frequency and stimulus quality. In contrast, recent reanalyses of previously published data from experiments that, unlike the Masson and Kliegl experiments, did not include semantic priming as a factor, found no evidence for modulation of additive effects of frequency and stimulus quality by trial history (Balota, Aschenbrenner, & Yap, Journal of Experimental Psychology: Learning, Memory, and Cognition, 39, 1563–1571, 2013; O’Malley & Besner, Journal of Experimental Psychology: Learning, Memory, and Cognition, 34, 1400–1411, 2013). We report two experiments that included semantic priming as a factor and that attempted to replicate the modulatory effects found by Masson and Kliegl. In neither experiment was additivity of frequency and stimulus quality modulated by trial history, converging with the findings reported by Balota et al. and O’Malley and Besner. Other modulatory influences of trial history, however, were replicated in the new experiments and reflect potential trial-by-trial alterations in decision processes.     

Efficient <Emphasis Type="Italic">n</Emphasis>-gram analysis in R with <Literal>cmscu</Literal>

We present a new R package, cmscu, which implements a Count-Min-Sketch with conservative updating (Cormode and Muthukrishnan Journal of Algorithms, 55(1), 58–75, 2005), and its application to n-gram analyses (Goyal et al. 2012). By writing the core implementation in C++ and exposing it to R via Rcpp, we are able to provide a memory-efficient, high-throughput, and easy-to-use library. As a proof of concept, we implemented the computationally challenging (Heafield et al. 2013) modified Kneser–Ney n-gram smoothing algorithm using cmscu as the querying engine. We then explore information density measures (Jaeger Cognitive Psychology, 61(1), 23–62, 2010) from n-gram frequencies (for n=2,3) derived from a corpus of over 2.2 million reviews provided by a Yelp, Inc. dataset. We demonstrate that these text data are at a scale beyond the reach of other more common, more general-purpose libraries available through CRAN. Using the cmscu library and the smoothing implementation, we find a positive relationship between review information density and reader review ratings. We end by highlighting the important use of new efficient tools to explore behavioral phenomena in large, relatively noisy data sets.     

Bayesian inference with Stan: A tutorial on adding custom distributions

When evaluating cognitive models based on fits to observed data (or, really, any model that has free parameters), parameter estimation is critically important. Traditional techniques like hill climbing by minimizing or maximizing a fit statistic often result in point estimates. Bayesian approaches instead estimate parameters as posterior probability distributions, and thus naturally account for the uncertainty associated with parameter estimation; Bayesian approaches also offer powerful and principled methods for model comparison. Although software applications such as WinBUGS (Lunn, Thomas, Best, & Spiegelhalter, Statistics and Computing, 10, 325–337, 2000) and JAGS (Plummer, 2003) provide “turnkey”-style packages for Bayesian inference, they can be inefficient when dealing with models whose parameters are correlated, which is often the case for cognitive models, and they can impose significant technical barriers to adding custom distributions, which is often necessary when implementing cognitive models within a Bayesian framework. A recently developed software package called Stan (Stan Development Team, 2015) can solve both problems, as well as provide a turnkey solution to Bayesian inference. We present a tutorial on how to use Stan and how to add custom distributions to it, with an example using the linear ballistic accumulator model (Brown & Heathcote, Cognitive Psychology, 57, 153–178. doi: 10.1016/j.cogpsych.2007.12.002, 2008).     

An examination of the processing capacity of features in the Thatcher illusion

Detection of the Thatcher illusion (Thompson, Perception, 9:483–484, 1980) is widely upheld as being dependent on configural processing (e.g., Bartlett & Searcy, Cognitive Psychology, 25:281–316, 1993; Boutsen, Humphreys, Praamstra, & Warbrick, NeuroImage, 32:352–367, 2006; Donnelly & Hadwin, Visual Cognition, 10:1001–1017, 2003; Leder & Bruce, Quarterly Journal of Experimental Psychology, 53A:513–536, 2000; Lewis, Perception, 30:769–774, 2001; Maurer, Grand, & Mondloch, Trends in Cognitive Sciences, 6:255–260, 2002; Stürzel & Spillmann, Perception, 29:937–942, 2000). Given that supercapacity processing accompanies configural processing (see Wenger & Townsend, 2001), supercapacity processing should occur in the processing of Thatcherised upright faces. The purpose of this study was to test for evidence that the grotesqueness of upright Thatcherised faces results from supercapacity processing. Two tasks were employed: categorisation of a single face as odd or normal, and a same/different task for sequentially presented faces. The stimuli were typical faces, partially Thatcherised faces (either eyes or mouth inverted) and fully Thatcherised faces. All of the faces were presented upright. The data from both experiments were analysed using mean response times and a number of capacity measures (capacity coefficient, the Miller and Grice inequalities, and the proportional-hazards ratio). The results of both experiments demonstrated some evidence of a redundancy gain for the redundant-target condition over the single-target condition, especially in the response times in Experiment 1. However, there was very limited evidence, in either experiment, that the redundancy gains resulted from supercapacity processing. We concluded that the oddity signalled by inversion of eyes and mouths does not arise from positive interdependencies between these features.