Merry-or-Worry,Illusion

Quite some years ago, a Krups mechanical weighing scale 834-20-10, type K11 (made in France) happened to find its place in our household. It had a pleasant text on the pedestal: ‘Have a nice day!’ Indeed, the text was accompanied by a sunny-looking face with a smile. However, suddenly one day, the undersigned noticed with a bit of astonishment that the greeting in the text and the habitual view of the merry chap was toned down by the worrisome expression of the same face of his. By that time, as a relatively young perception researcher, I was already well versed in knowing many standard multistable figures such as Necker cube, Rubin's vase/faces, the rabbit/duck and the wife/mother-in-law figures (popularized, respectively, by Joseph Jastrow and Edwin Boring). (For pertinent reviews and illustrations, see, e.g., Bach, 1997; Fisher, 1968; Long & Toppino, 2004 or Toppino & Long, 2015.) Yet, the Krups weighing scale ambiguous figure fascinated me, especially because, in addition to the typical perceptual alternations between the merely two visual conscious interpretations, this ambiguous image caused also an alternation between different directly experienced emotional feelings. Ambiguity of facial expression of affect was clearly present.

However, I have not tried the effect of this illusory fluctuation of perceived emotion on anybody else. Maybe this is somehow related to the function of such weighing scales, who knows. Now, in the illustration drawn here (Figure 1) I have finally tried to depict such an image by myself from memory, based on what I have experienced in the Krups artefact pedestal face (accompanied by subjective fascination, but this is no wonder – Stevanov et al., 2012).

OPEN IN VIEWER

Figure 1.

The schematic face allowing illusory alternation between perceptual impressions of a worrisome and a happy-looking (merry) face.

To experience the ambiguity (e.g., reversibility, bistability, multistability, and fluctuation), you are invited to fixate your gaze in the centre of the face and keep fixation. Sooner or later, the perceived facial expression changes and can alternate between the merry-looking and worrisome-looking variants.

The capacity of the same face-depicting object to carry variable expressive qualities is not something unforeseen. For example, popular Internet sources are full of pictures of faces which, when turned upside down, become perceived as different faces with different expressions (e.g., many examples created by artists – https://www.pinterest.com/pin/794040978020299971/). Among the most enjoyable examples, we meet pictures by Rex Whistler. By having such an image rotated by 180°, facial expression changes (for illustration, see Wade & Nekes, 2005). No less known is the Thatcher illusion (Thompson, 1980): when basic local elements of a face such as mouth and eyes are rotated upside down within the upright full facial image, the impression is grotesque (and perhaps even frightening for someone). If you rotate that weird image upside down, the grotesque quality is not perceived anymore. However, in the above-mentioned illusions, the mutual spatial relations between the elements of the face are not changed. Only whole-image and local element inversions are physically introduced. On the other hand, the present merry-or-worry face can be held always in the canonical orientation and the orientation of local elements need not be changed for the illusion to be experienced. Spatial interrelations of the facial elements in the objective image are also invariant. What keeps the illusion working? Subjective reinterpretation of the meaning of the image elements appears crucial.

To argue for this simple cause, two disambiguating pictures are shown (Figure 2).

OPEN IN VIEWER

Figure 2.

Disambiguated faces: Merry to the left and worry to the right. The expressive content of the mouth element implicates the affective content.

In the face on the left (Figure 2), an upwards concave mouth typically associated with happiness is kept, but the downwards concave mouth associated with sadness/worrying is deleted – the character is only merry (happy). In the face to the right, only the worry-implicating shape of mouth is present and the chap, respectively, feels just worrying.

It is worth noting that in the well-known ambiguous figures, the same elements become ‘owned’ by different object entities when interpretation reverses (e.g., the eye of an old lady becomes the ear of a young lady who is a different object entity). In the case of the Merry-or-Worry (Figure 1), the object entity is the same, but its features are reinterpreted. As reinterpreted, the element becomes in a sense also spatially translated. What is similar though is swapping or exchange between the meaning of the element: to experience the illusion of change, the merry-face mouth has to be reinterpreted as a moustache and the mouth of the worrying face as a little wrinkle in the chin area.

For the present author, the usefulness of this type of ambiguous figure for more technical and precise studies of object and feature (predictive coding) perception mechanisms is acknowledged. For example, consider feature-swapping versus object-token-invariance stimuli used against typical object-swapping stimuli combined with brain imaging for neural markers of conscious perception. (Using expressively multistable facial stimuli with cognitive neuroscience methods already has its precedents – for example, Carbon et al., 2005; Psalta et al., 2014.) The only ‘experiment’ carried out so far with the merry-or-worry stimulus involves just one observer (T.B.) and one ‘apparatus’ (hands of T.B.). I rotated the image gradually from 0° up to 180° (this ‘powerful’ exploratory experiment had 10 trials). Interestingly, the phenomenal multistability quite abruptly went difficult to keep after about 100°–120° rotation. Curiously, this multistability threshold is very similar to what Stürzel and Spillmann (2000) found for the tolerance of the Thatcher illusion. I leave the theoretical interpretation of this for future though.

To hypothesize and conclude: for the merry-or-worry face introduced here, the subjective, phenomenally experienced content related grouping (Strüber & Stadler, 1999) and/or imaginary top-down accentuation (Pinna et al., 2018; Verstijnen & Wagemans, 2004) of (some of) the basic elements of the face determines which variant of expression you perceive. If the element right below the nose is interpreted as the mouth, the face looks merry or worriless. On the other hand, if that element is interpreted as the moustache and the lowermost element as mouth, the face turns worrying. The nice side of this is that it turns to smile again.