Basically what they discovered was that if you have subjects do something like bite on a pencil (so that they come up with a smile, of sorts) or just keep repeating the high front vowel /i/ that has that
articulatory setting while they watch a cartoon, they tend to see things as more amusing. If, on the other hand, you have them stick the end of that pencil in their mouth so that they develop an extreme pucker, or keep repeating the vowel /o/, they tend to see things as less amusing.
So? It has been known for decades that vowels do have phonaesthetic qualities. (See several previous blog posts.) The question has always been . . . but why? The conclusion: Because of what the facial muscles are doing while the vowel is articulated, especially as it relates to non-lexical (non word) emotional utterances. Could be, but they should have also tossed in some controls, some other vowels, too, such as having subjects use a mid, front unrounded vowel such as /ae/, as in "Bad!"-- or a high front rounded vowel, such as /ü/, as "Uber," the web-based taxi service, or a high back unrounded vowel.
As much as I like the haptic pencil technique, which I use myself occasionally (using coffee stirs, however) for anchoring lip position with those vowels and others, there is obviously more going on here, such as the phonaesthetic qualities of the visual field. Also consider the fact that the researchers appear to be ethnically German, perhaps seriously compromising their ability to even perceive "amusing" in the first place, conducing them into that interpretation of the results.
Nonetheless, an interesting and possibly useful study for us, more than mere "lip" service, to be sure.