Aug. Serm. 378 distinguished clearly between arrabo and pignus: quando datur pignus reddit homo quod accepit re completa propter quam pignus accepit, arra autem quando datur non recipitur sed super additur ut impleatur. The same distinction is drawn by Isid. Orig. v 25.20. Especially significant is the assumption by Gellius, that there was a usual distinction known to everybody, and his consequent explanation of why Quadrigarius used arrabo where pignus would be expected (xvii 2.21): Cum tantus, inquit, arrabo penes Samnites populi Romani esset. Arrabonem dixit sescentos obsides et id maluit quam pignus dicere, quoniam vis huius vocabuli in ea sententia grantor acriorque est. Quadrigarius used arrabo, I take this to mean, because it expressed more keenly the bitterness felt by the Romans at being forced to leave their fellow-citizens in pawn. (If this is so arrabo would already be in sordidis verbis, though Gellius ascribes this colour of the word to his own times as against those of Quadrigarius.)