The words we use have no inherent capacity to evoke negative or positive affect.  Instead, how we appraise, reappraise, attend to, and reflect on those words determines our affective response (e.g., [Gross, 1998](http://doi.apa.org/psycinfo/1998-10496-003); [Siemer, Mauss, & Gross, 2007](http://psycnet.apa.org/journals/emo/7/3/592/)).

For example, you could tell one person "You are stupid" and he/she might become extremely upset.  You could tell another person the same thing, and he/she would have no affective response.  Similarly, telling someone "You are nice" could make someone very happy or not change his/her mood at all.

So how can we predict who will respond most intensely? [Philippe Verduyn](https://scholar.google.com/citations?user=gqi5hA8AAAAJ&hl=en&oi=ao)  and colleagues ([2009](http://psycnet.apa.org/journals/emo/9/1/83/); [2011](http://psycnet.apa.org/journals/emo/11/1/20/)) have looked at predictors of greater emotion response duration, which include:

 - The salience of the emotion-eliciting event
 - How you conceptualize your positive/negative affect (joy lasts longer than anger, which lasts longer than fear)
 - Intensity of the emotion at onset 
 - Physical reappearances of the emotion-eliciting stimulus
 - Stimulus-related cognitions with the same valence as the emotion
 - Social sharing (consistent with [Bernard Rimé's](https://scholar.google.com/citations?user=qE0UjjsAAAAJ&hl=en&oi=ao) work)

How one chooses to regulate their emotions in response to an insult or compliment would also influence emotion reactivity and duration, including tendency to distract oneself from the stimulus ([Sheppes et al., 2014](http://psycnet.apa.org/psycarticles/2012-30627-001)) or skill in reappraising the situation (e.g., [McRae et al., 2012](http://www.sciencedirect.com/science/article/pii/S0092656611001437)).

We would also expect depressive rumination (see [Susan Nolen-Hoeksema's work](https://scholar.google.com/scholar?hl=en&q=susan+nolen-hoeksema&btnG=&as_sdt=1%2C44&as_sdtp=)) and positive rumination or savoring (e.g., [Feldman, Joormann, & Johnson, 2008](http://link.springer.com/article/10.1007/s10608-006-9083-0); [Quoidbach, Berry, Hansenne, & Mikolajczak, 2010](http://www.sciencedirect.com/science/article/pii/S0191886910001820)) to predict intensity of responses to insults and compliments, respectively.

## Conclusion ##

So the point is that there are no special words or phrases that will reliably evoke some level of positive affect.  Instead, the emotional intensity evoked by a stimulus depends on the salience of the stimulus, proximity to the stimulus, and individual differences in emotional responding (i.e., reactivity and regulation; see [Gross & Barrett, 2011](http://emr.sagepub.com/content/3/1/8.short)).  If you take these into account, you can devise more powerful insults and compliments.