180 Gréte Dalmi
In Hungarian, where free, genericity-inducing one vs. bound variable one are real-
ized as two lexically distinct forms (az emberGN ‘the man’ and proGN), it is not surpris-
ing that the PRO subject of infinitival clauses also displays such duality. Compare (65)
and (66):
Free pronominal PROarb
(65) Nem volna lehetséges (*az ember-nek) [ten-ni
not would_be possible (the man-dat) do-inf
PROarb valami-t]?
something-acc
‘Wouldn’t it be possible (*for one) [to do something]?’
Bound variable PROGN
(66) Élvezetes volna proGN [eb-ben
enjoyable cop.cond 3 sg (for one) this-iness
a park-ban sétálni PROGN].
the park-iness walk-inf
‘It would be enjoyable (for one) to walk in this park.’
As is clear from (65), PROarb does not accept a generic inclusive controller; in fact, it
does not need any antecedent in order to obtain the generic interpretation. Moreover,
PROarb shows correlation with generic exclusive people in that it interacts with quan-
tifiers, while PROGN does not. This can be tested by adding a quantificational adverb
like néha ‘sometimes’:
(65’) Nem volna lehetséges [néha
not would_be possible sometimes
ten-ni is PROarb valami-t]?
do-inf also something-acc
‘Wouldn’t it be possible [to do something as well, sometimes]?’
(66’) *Élvezetes volna proGN [néha
enjoyable cop.cond 3 sg (for one) sometimes
eb-ben a park-ban sétálni PROGN].
this-iness the park-iness walk-inf
‘It would enjoyable (for one) to walk in this park sometimes.’
(65’) allows for the interpretation ‘aren’t there cases when it would be possible (for
people) to do something.....’, while this interpretation is not available in (66’): *’there
are cases when it would be enjoyable (for one) to walk in this park’.
As is clear from (66), generic inclusive PROGN can have a non-nominative null
antecedent (for non-nominative subjects in general see Sigurðsson 2000, 2002 , 2006 ,
2010 and the references cited therein; for dative subjects in Hungarian see Dalmi 1983,