Abstract
In this paper we will propose a simple linguistic approach to the Knobe effect, or the moral asymmetry of intention attribution in general, which is just to ask the felicity judgments on the relevant sentences without any vignette at all. Through this approach we were in fact able to reproduce the (quasi-) Knobe effects in different languages (English and Japanese), with large effect sizes. We shall defend the significance of this simple approach by arguing that our approach and its results not only tell interesting facts about the concept of intentional action, but also show the existence of the linguistic default, which requires independent investigation. We will then argue that, despite the recent view on experimental philosophy by Knobe himself, there is a legitimate role of the empirical study of concepts in the investigations of cognitive processes in mainstream experimental philosophy, which suggests a broadly supplementary picture of experimental philosophy.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Note that Strickland et al. (2015) reports that syntax affects our intentionality judgements. Thus, in this sense language is relevant to the intention-attribution. But we are here concerned with the relevance of the semantic aspect or concepts captured by the relevant expressions.
On the other hand, experimental philosophy of the negative program denies the reliance on intuitions in philosophy in general, or in its weaker form, the reliance on intuitions without any empirical support. Weinberg (2016) takes the latter as the negative program of experimental philosophy. If so, it corresponds to Experimental Restrictionism of Nadelhoffer and Nahmias (2007). See Sect. 5.2 of Weinberg (2016) for more on the negative program.
Nadelhoffer and Nahmias (2007) called such a project Experimental Analysis, and Machery (2016, Sect. 33.2) called it (in philosophy of science) experimental conceptual analysis. But note that, the supplementary picture can also be shared by armchair philosophers, and moreover, pace Knobe (see below), whether a paper or the result reported there is a contribution to conceptual analysis or not is not an inherent property of the paper or the result, but is rather a matter of how we use such result. Thus, K. Mortensen and J. Nagel use various results concerning, e.g., free will (Nahmias et al. 2006; Nichols and Knobe 2007; Murray and Nahmias 2014), phenomenal state ascription (Buckwalter and Phelan 2014), and the pain paradox (Reuter et al. 2014), to show that “empirical work can extend and enhance the reach of traditional philosophical theorizing” (Mortensen and Nagel 2016, p. 64). In general, what Mortensen and Nagel say in Sect. 4.5 of their 2016 (especially about what is called the “neutral” project there) is closely related to what I argue in this section.
Though in his (2003a), Knobe heavily emphasizes implications of his findings for the concept of intentional action as their significance.
But note the caveat concerning this categorization in footnote 4 above.
So, if Knobe thinks that contributing to conceptual analysis is a morally good thing, he himself might have been affected by the Knobe effect.
It seems that, especially if there are also influences from concepts to cognitive processes (see below), the virtue of the simplicity of the theory can no more be expected for the theory of cognitive process than the one based on conceptual analysis. No doubt Knobe proposes an independent criterion of simplicity of explanation of a particular cognitive process relative to a complex general theory (ibid. sec. VII). But, then, the same criterion can and should be applied to conceptual analysis. Note, however, it is in fact acceptable for us to admit that no simple theory can be expected in either area. In such a case, Wittgensteinian quietism may prevail, and all the philosophical problems are to be dissolved, rather than solved. This will be done by understanding better and more perspicuously our own form of life. Indeed, after we obtaining the results about the particular cognitive processes, we have the right, or even the duty, to ask why we have such (rather than other) cognitive processes, which should be answered by our social facts lying outside of the head of individuals.
Assumption 1 invites a variant of the notorious problem of drawing a clear line between meaning and belief, discussed by Donald Davidson (for his claim of the interdependence of belief and meaning, see Davidson 1973, p. 134, 1985, p. 158). If Davidson is right, concepts and cognitive processes should be likewise intertwined. This is also a matter of whether the relevant norms (governing the use of the relevant term) are at the linguistic level or psychological level. But in either case, at least for the externalist about meaning and content they exist outside of the head of individuals as social conventions.
Note that the mutual influence relation has been exactly what is implied by the moral asymmetry in question. Intention-attributions, causal judgments, and others do have direct consequences on our moral judgments. What Knobe and others have revealed was that there is also the influence of the opposite direction.
For example, Japanese has two knowledge verbs “shitte-iru” and “wakatte-iru”, and it has been reported that they express quite different concepts even when used for propositional knowledge (Mizumoto forthcoming). If so, this should support pluralism of the concept of knowledge.
Note also that, Davidson’s argument for the interdependence of meaning and belief was presented in the cross-linguistic context, or that of radical interpretation.
There is in fact another adverb, “koi ni (故意に)”, which lies just in-between “itoteki ni” and “wazato”, in the sense that it is not newly created recently like “itoteki ni”, but is still a formal, rather than colloquial, expression. For comparison, however, we chose the two opposite ends of the continuum.
It gives four old Japanese uses, only one of which is the present sense of “intentionally”. Others are not in use or rarely found in contemporary Japanese.
In the actual sentences presented to participants, “X” was replaced by “John” in the English survey, and “Taro” in the Japanese surveys.
Similarly, “X” was replaced as described in the last footnote, and “Y” was replaced by “Betty” in the English surveys and “Hanako” in the Japanese surveys.
In Japanese, (1) Tadashiku shizen, (2) Machigai de wa nai ga, hushizen, and (3) Machigai.
Note that the decision to use this dichotomy is obviously post hoc. However, this is not ad hoc like the dichotomy “not wrong but unnatural” or otherwise (correct or wrong). We did not use the dichotomy correct (natural or unnatural) or “wrong” for it obviously has no difference between Neutral, Bad, and Good, which is why this is a post hoc analysis. See also the next footnote.
One might think that, we should conduct ordinal regression analysis here. However, it is questionable that we can take “correct and natural”, “unnatural”, and “wrong” as ordered in this way, rather than categorical distinction. Besides, the present dichotomy makes perfect sense, as pointed out in the last footnote. Indeed, ideally, we could use the interaction model N × (TUMBLE + BREAK + IGNORE) + B × (STEAL + HARM + KILL) + G × (IMPROVE + SAVE). We avoided these methods just for the sake of simplicity, especially given the already extremely good fit of the present models (N + B + G).
The latter two models themselves were significant (the base model: χ2 (7) = 323.5, p < 0.0001, the model (N or B) + G: χ2 (1) = 252.3, p < 0.0001).
Again, the latter two models themselves were significantly better fit than the null hypothesis (the base model: χ2 (7) = 608.0, p < 0.0001, the model (N or B) + G: χ2 (1) = 299.0, p < 0.0001).
In the case of TUMBLE, the infelicity may come from the fact that it is simply not so easy to intentionally (or itoteki ni) tumble unless, for example, you are a trained comedian. In the case of KILL, there may be something special about the murder over and above its being morally bad. For example, there may be some special (semantic or pragmatic) connection between “wazato” and killing, so that “wazato” in KILL is especially infelicitous.
The difference between “itoteki ni” and “wazato” in HARM failed to be significant (p = 0.16, two-tailed Chi square test), but otherwise all other differences in both HARM and IMPROVE turned out to be strongly significant (p < 0.0001).
One may therefore think here that the use of “itoteki ni” is also affected by “wazato”, as well as “intentionally”. To show this more rigorously might require surveys with within-subject design using bilinguals. But the point here is that, if this is a correct account, the effect in a survey with some term might be a result of the effect of some concept captured by some other term if a (possible) survey with the latter term has a larger effect size than the original survey (we shall come back to this issue in the final section).
Rejecting such a possibility may be based on a principle that connects the correct application of a term with the truth of a sentence involving it, like “intentionally” is correctly applied to S’s φ-ing if and only if “S intentionally φs” is true, which seems plausible enough.
Alas, it turned out that there is a large difference in the use of truth predicates in different languages due to the difference of sensitivity to moral-political factor in the utterance (**** manuscript). But if so, the truth-conditional content itself should also vary from language to language.
According to Bach (1997), the three major distinctions between semantic/pragmatic distinction are: 1. linguistic meaning versus use, 2. truth-condition versus non-truth-conditional meaning, 3. context independence versus context dependence. Though all of them beg the question against some view of semantics, the present distinction is especially relevant to 3.
Note that, unlike conversational implicature, conventional implicature is semantic meaning (cf. Potts 2007, Sec. 3).
But if so, this suggests an unexpected return of the pragmatic account of the Knobe effect for English speakers (Adams and Steadman 2004a, b; Driver 2008a, b), which is compatible with the psychological account but has long been regarded as refuted (cf. Knobe 2010, Sect. 4.2). We need to take the pragmatic account seriously again at least in the case of English speakers.
See also the same implication of two Japanese knowledge verbs, mentioned at footnote 13.
I would like to thank Josh Knobe for suggesting me to clarify this point.
Compare this with the notion of default in Sect. 5.2 of Knobe (2010). It is ordinary people’s reference point based on their conception of alternative possibility in each case, relative to which the agent’s con/pro attitude is assessed. This is supposed to be determined by people’s psychology, rather than linguistic norms (like our linguistic default).
If so one might think that it could be investigated by traditional conceptual analysis through the method of cases. But even if that is true, mere conceptual analysis cannot determine the degree of the felicity of the kind we report here.
But if the context of use can strengthen the moral asymmetry already existing at the linguistic level, it may also reduce or eliminate the asymmetry. It might be possible to move the place of the linguistic default and render people judge SAVE and IMPROVE to be correct and natural even in the “wazato” survey, by providing an appropriate background story. If this happens, then that part of the effect is truly a psychological process over and above the linguistic/conceptual effect.
See footnote 26 above.
References
Adams, F., & Steadman, A. (2004a). Intentional action in ordinary language: Core concept or pragmatic understanding? Analysis, 64, 173–181.
Adams, F., & Steadman, A. (2004b). Intentional actions and moral considerations: Still pragmatic. Analysis, 64, 268–276.
Alexander, J. (2012). Experimental Philosophy: An Introduction. Cambridge: Polity Press.
Bach, K. (1997). The semantics-pragmatics distinction: What it is and why it matters. Pragmatik, 33–50.
Buckwalter, W., & Phelan, M. (2014). Phenomenal Consciousness Disembodied. In Justin Sytsma (Ed.), Advances in Experimental Philosophy of Mind (pp. 45–73). New York: Bloomsbury.
Cushman, F. (2014). What scientific idea is ready for retirement? Big effects have big explanations. Retrieved from http://edge.org/response-detail/25508.
Davidson, D. (1973). Radical interpretation. In Inquiries into truth and interpretation (pp. 125–140). Oxford: Clarendon Press (1985).
Davidson, D. (1975). Thought and talk. In Inquiries into truth and interpretation (pp. 125–140). Oxford: Clarendon Press (1985).
Driver, J. (2008a). Attributions of causation and moral responsibility. In W. Sinnott-Amstrong (Ed.), Moral psychology: The cognitive science of morality—Intuition and diversity (Vol. 2, pp. 423–439). Cambridge, MA: MIT Press.
Driver, J. (2008b). Kinds of norms and legal causation: Reply to Knobe and Fraser and Deigh. In W. Sinnott-Amstrong (Ed.), Moral psychology: The cognitive science of morality—Intuition and diversity (Vol. 2, pp. 459–461). Cambridge, MA: MIT Press.
Knobe, J. (2003a). Intentional action and side effects in ordinary language. Analysis, 63(279), 190–194.
Knobe, J. (2003b). Intentional action in folk psychology: An experimental investigation. Philosophical Psychology, 16(2), 309–324.
Knobe, J. (2007). Experimental philosophy and philosophical significance. Philosophical Explorations, 10, 119–122.
Knobe, J. (2010). Person as scientist, person as moralist. Behavioral and Brain Sciences, 33(04), 315–329.
Knobe, J. (2016). Experimental philosophy is cognitive science. In Sytsma, J., & Buckwalter, W. (Eds.), A companion to experimental philosophy (pp. 78–96).
Knobe, J., & Nichols, S. (2008). An experimental philosophy manifesto. In J. Knobe & S. Nichols (Eds.), experimental philosophy (pp. 3–14). New York: Oxford University Press.
Machery, E. (2016). Experimental philosophy of science. In Sytsma, J., & Buckwalter, W. (Eds.), A companion to experimental philosophy (pp. 475–490).
Mizumoto, M. (forthcoming). “Know” and Japanese counterparts: “Shitte-iru” and “Wakatte-iru”. In Mizumoto, M., Stich, S., McCready, E., & Stanley J. (Eds.), Epistemology for the rest of the world. Oxford: Oxford University Press.
Mortensen, K. & Nagel, J. (2016). Armchair-friendly experimental philosophy. In Sytsma, J., & Buckwalter, W. (Eds.), A companion to experimental philosophy (pp. 53–70).
Murray, D., & Nahmias, E. (2014). Explaining Away Incompatibilist Intuitions. Philosophy and Phenomenological Research, 88, 434–467.
Nadelhoffer, T. (2005). Skill, luck, control, and folk ascriptions of intentional action. Philosophical Psychology, 18, 343–354.
Nadelhoffer, T., & Nahmias, E. (2007). The past and future of experimental philosophy. Philosophical Explorations, 10(2), 123–149.
Nahmias, Eddy, Morris, Stephen G., Nadelhoffer, Thomas, & Turner, Jason. (2006). Is Incompatibilism Intuitive? Philosophy and Phenomenological Research, 73, 28–53.
Nichols, S., & Knobe, J. (2007). Moral responsibility and Determinism: The cognitive science of folk intuitions. Nous, 41, 663–685.
Phillips, J., Luguri, J. B., & Knobe, J. (2015). Unifying morality’s influence on non-moral judgments: The relevance of alternative possibilities. Cognition, 145, 30–42.
Potts, C. (2007). Into the Conventional-Implicature Dimension. Philosophy compass, 2(4), 665–679.
Reuter, K., Phillips, D., & Sytsma, J. (2014). Hallucinating Pain. In Justin Sytsma (Ed.), Advances in Experimental Philosophy of Mind (pp. 75–99). New York: Bloomsbury.
Robinson, B., Stey, P., & Alfano, M. (2015). Reversing the side-effect effect: The power of salient norms. Philosophical Studies, 172(1), 177–206.
Stich, S., & Tobia, K. (2016). Experimental philosophy and the philosophical tradition. In Sytsma, J., & Buckwalter, W. (Eds.), A companion to experimental philosophy (pp. 5–21).
Strickland, B., Fisher, M., Knobe, J., & Keil, F. (2015). Syntax and intentionality: An automatic link between between language and theory-of-mind. Cognition, 133(1), 249–261.
Sytsma, J., & Buckwalter, W. (Eds.). (2016). A companion to experimental philosophy. Hoboken: Wiley.
Sytsma, J., & Machery, E. (2013). Experimental philosophy. In Hancock, P. A., & Hancock, G. M. (Eds.), Encyclopedia of philosophy and the social sciences (pp. 319–321).
金光林(2005), 近現代の中国語、韓国 ・ 朝鮮語における日本語の影響—日本の漢字語の移入を中心に—新潟産業大学人文学部紀要 第17号, 111–238.
Weinberg, J. M. (2016). Going positive by going negative: On keeping X-Phi relevant and dangerous. In Sytsma, J., & Buckwalter, W. (Eds.) (pp. 71–86).
Acknowledgements
I would like to thank Yoshihisa Kashima for encouragement and Joshua Knobe for kindly reading and commenting on an earlier manuscript. This work was supported by J. S. P. S Kakenhi (C) Grant No. JP26370010.
Author information
Authors and Affiliations
Corresponding author
Appendix: Japanese data of the survey with the original Chairman case
Appendix: Japanese data of the survey with the original Chairman case
We omit here the demographic details of the participants, but they were recruited through Lancer with exactly the same procedures as in the main text. The results are summarized in the following table with the data of “intentionally” taken from Knobe (2003a) for comparison.
HARM (%) | IMPROVE (%) | |
|---|---|---|
Itoteki ni | 86.0 (N = 50) | 11.8 (N = 51) |
Wazato | 76.8 (N = 56) | 12.5 (N = 56) |
Intentionally | 82 (N = 39) | 23 (N = 39) |
As for HARM, there was no significant difference (Fisher’s exact test) between the results of “itoteki ni” and “wazato”. The statistical power (1-β) of detecting the difference between these two in HARM for effect size of 0.3 and this sample size (N = 106) was 0.87.
In the case of IMPROVE, even if we doubled the sample size of “intentionally” to 78 (while keeping the percentage constant, which we assumed 9/39), neither the difference with “itoteki ni” nor “wazato” was significant (Fisher’s exact test).
Overall, we found no significant difference between three adverbs (even after doubling the sample size of “intentionally”) either in HARM or IMPROVE (Chi square test), which was predictable as discussed in Sect. 4.
Rights and permissions
About this article
Cite this article
Mizumoto, M. A simple linguistic approach to the Knobe effect, or the Knobe effect without any vignette. Philos Stud 175, 1613–1630 (2018). https://doi.org/10.1007/s11098-017-0926-1
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11098-017-0926-1