Jägermonsterizer

More mad science from the Foglios and their fans: Here's a Jäger-accented paragraph from the background section of my thesis:

Hevink a direck evaluation measure (as PARSEVAL provides) hallows to pursue dis direction uf research. De PCFG (und leksicalized PCFG) independence assumptions reqvire dat de overall probability assigned to a parse kundidate must be built up over tree-local decisions. But [Kollins, Johnson] take anodder approach: de n-best kundidates selected by de PCFG may be treated as a set uf kundidates for discriminatiff rerankink, vith a vider variety uf features.