The mid-C13th satire De palpone by the Franciscan schoolmaster Walter of Wimborne includes this stanza (§64, spelling modified):
Palpo sententiae favet utrilibet,
gratus quibuslibet quia qualislibet;
contingens etenim est ad utrumlibet,
vel impossibile quod infert quidlibet.
The first couplet is straightforward enough:
The flatterer favours whichever opinion,
he pleases whoever 'cause he's a chameleon;
For the second couplet, George Rigg suggests in his 1978 edition: ‘“He is contingent on (depends on) either side, or on whatever impossible inference is made.” That is, the flatterer is like a conclusion in logic, dependent on the preceding premise, however impossible it may be. AB agree on quidlibet, but quilibet would be better.’
As I suspect anyone reading this will have noticed, this interpretation is faulty. The contingens line has nothing to do with dependence, but imputes to the flatterer an indeterminate attitude towards pairs of contradictory propositions. And the proposed emendation would break the allusion in the last line to the logical principle often termed ex falso quodlibet. If the text is correct, then, we may translate:
for he's the contingent towards either side,
or else the impossible all things implying.
Here I have construed Wimborne as metaphorically identifying the flatterer with things that have certain modal properties. If instead he were metaphorically ascribing these modal properties directly to the flatterer, we would expect impossibilis instead of impossibile, in which case (understanding quod as ‘because’) we could translate:
for he is contingent towards either side —
or rather impossible, all things implied.
I know nothing about mediaeval poetry, so for all I know this second suggestion may be metrically untenable. But the hybrid contingens ad impossibile construction that Rigg discerns here is very odd – and I flatter myself that both of my suggestions are more amusing.