This paper explores the referential properties of the one-word reciprocal form toisiaan 'other-PL-Px' and the doubled reciprocal form toinen toisiaan 'other other-PL-Px' in Finnish. This research builds on work presented in Kaiser (to appear), where the author argued that the distinction between strong vs. weak reciprocity is not sufficient to explain the division of labor between the one-word form and the doubled form, and hypothesized that the doubled form is less tolerant of exceptions than one one-word form. Building on work by Beck (2000, 2001) and Brisson (1998, 2003), the author suggested that toinen in the doubled form increases the discourse-prominence of good-fit covers. The present paper extends this research by investigating the consequences of using the doubled form in contexts where good-fit readings are already favored by other factors and by comparing 'zero reciprocal' verbs to the one-word and the doubled forms. The author notes that using the doubled form in a context where a good-fit reading is already preferred does not result in infelicity and may in fact be interpreted as conveying some additional information, possibly regarding the distinguishability of the events. Furthermore, the author explores the possibility that the doubled form's association with exceptionless interpretations is due not to toinen increasing the salience of good-fit covers but rather to general reasoning/inferences regarding prolixity vs. brevity of referential form, with the exceptionless preference of the doubled form merely being one type of 'marked meaning' conveyed by the choice of a prolix form. The author discusses data from the zero reciprocal paradigm that present a challenge for this type of approach.