Although we think that the development of our field benefits from critical discussion, the scientific contribution of the current target article escaped us. It is neither a theoretical article, as it lacks a theory, nor is it a good review article, because it is biased and selective. In our commentary, we first focus on the lack of theory and subsequently on the misleading treatment of the literature.

In our view, the goal to incorporate processes that we are not conscious of in our theories was at least partly sparked by the wish to make our field more mature. Making psychology compatible with philosophy and neuroscience and getting rid of homunculi was, and is, a worthwhile endeavor. This goal led to a burgeoning number of theories and models on unconscious processes – in which there usually is a role for conscious processes, though a modest one – and critical responses from colleagues who maintain that all important psychological processes are guided by consciousness.

Newell & Shanks (N&S) defend an extreme version of the latter position, but fail to provide a reasonable alternative. If conscious thoughts do not have unconscious precursors, as N&S seem to believe, then where do they come from? Unless one believes that consciousness freely hovers in the air, or is in direct contact with the gods, claiming that psychological processes start in consciousness without further ado does not make sense. The house N&S try to build stands on scientific quicksand.

N&S's lack of theory is all the more disappointing in light of the recent scientific progress on the distinction between attention and consciousness, and on the relation between conscious and unconscious processes (Dehaene et al. Reference Dehaene, Changeux, Naccache, Sackur and Sergent2006; Dijksterhuis & Aarts Reference Dijksterhuis and Aarts2010; Koch & Tsuchiya, Reference Koch and Tsuchiya2006; Lamme, Reference Lamme2003; Wegner & Bargh, Reference Wegner, Bargh, Fiske and Gilbert1998). Unfortunately, they completely ignore this literature. Perhaps needless to say, stating “it's all starting in consciousness” may have been satisfactory to Descartes, but it doesn't work in the twenty-first century.

The target article also falls short of being a comprehensive review. It was mystifying to us why one would criticize the work by Libet (Reference Libet1985) without at least mentioning groundbreaking recent additions (e.g., Soon et al. Reference Soon, Brass, Heinze and Haynes2008), or why one would question prime-to-behavior effects on the basis of a few nonreplications without acknowledging the fact that such effects have been reported in well over a 100 papers.

Unavoidably, we read the section on unconscious-thought theory (UTT) attentively (sect. 3.2). We found the reasoning often flawed and the degree of cherry picking too extreme. However, we do agree with some observations of N&S. The strength of the initial evidence for unconscious-thought effects (UTEs) was indeed rather weak. The strong early claims such as the one “to leave decisions to the unconscious” (Dijksterhuis et al. Reference Dijksterhuis, Bos, Nordgren and van Baaren2006a) were, in retrospect, naïve. Finally, UTEs have been proven far from robust. It is indeed likely that there is a publication bias, but we all know this is a general problem, at least in psychology. Show us a psychological phenomenon studied in over 30 experiments, convince us there is no publication bias, and we will send an expensive bottle of wine.

That being said, the way N&S treat the UT literature does not do justice to the field. To begin with, they pre-emptively formulate some arbitrary inclusion criteria that allow them to discard dozens of experiments supporting UTT. On top of that, they ignore many papers fully compatible with their own criteria that do support UTT (e.g., Ham & van den Bos Reference Ham and van den Bos2010a; Reference Ham and van den Bos2010b; Reference Ham and van den Bos2011; Ham et al. Reference Ham, van den Bos and van Doorn2009; Handley & Runnion Reference Handley and Runnion2011; McMahon et al. Reference McMahon, Sparrow, Chatman and Riddle2011; Messner & Wänke Reference Messner and Wänke2011). An emphatic reader will understand that for people who have contributed to unconscious-thought research, reading this section was a tad discouraging.

N&S also suggest alternative explanations. They cite Newell et al. (Reference Newell, Wong, Cheung and Rakow2009), who asked participants to indicate their own attribute weights after they made their choice (instead of weights predetermined by the experimenter). Using these idiosyncratic a posteriori weights, choices in the conscious thought condition were just as good as in the unconscious thought condition. Obviously, participants are capable of generating post hoc weights that justify their previous choice. Usher et al. (Reference Usher, Russo, Weyers, Brauner and Zakay2011) measured participants' idiosyncratic attribute weights before the decision task and found the unconscious thought advantage predicted by UTT. N&S should have concluded that this completely refuted their idiosyncratic weights explanation, but they refrained from doing so because Usher et al.'s study did not satisfy their inclusion criteria (which are obviously irrelevant for his particular conclusion). Furthermore, they mention the possibility that participants make decisions before they can engage in unconscious thought, but fail to say that this explanation has already been ruled out in studies they conveniently disregarded (e.g., Bos et al. Reference Bos, Dijksterhuis and van Baaren2008).

We are optimistic about unconscious thought research, despite the clear limitations alluded to above. In a recent meta-analysis (Strick et al. Reference Strick, Dijksterhuis, Bos, Sjoerdma, van Baaren and Nordgren2011), moderators were found that led the UTE to be replicated with greater ease. Furthermore, new additions to the literature, such as a paper integrating UTT with fuzzy trace theory (Abadie et al. Reference Abadie, Waroquier and Terrier2013), and a paper reporting the first fMRI evidence for UT (Creswell et al. Reference Creswell, Bursley and Satpute2013) have appeared recently.

More generally, we strongly argue that consciousness and conscious decisions are best understood by their relation to unconscious processes. The most sensible approach to learn about conscious decisions is thus to consider higher cognitive operations as unconscious, and test what (if anything) consciousness adds rather than the other way around (e.g., Dijksterhuis & Aarts Reference Dijksterhuis and Aarts2010; van Gaal et al. Reference van Gaal, Lamme, Fahrenfort and Ridderinkhof2011; Zedelius et al. Reference Zedelius, Veling and Aarts2012). Although we surely agree that the road to progress in this field is rocky, focusing on consciousness without understanding its unconscious precursors is a dead end.