C88-2123 to make two remarks : ( i ) the feature percolation in example ( 6 ) could be done
E99-1029 actions , resulting in the need for feature percolation . For example , for the NP coordination
P90-1021 interesting phenomena , such as feature percolation . coordination , and many aspects
E89-1040 anaphoric relations . If we only feature percolation to encode anaphoric relations
P89-1034 subsequent concatena - tions . The feature percolation system underlies the partial
E99-1029 dispensed with . Note that , although feature percolation is eliminated when the EDOL hypothesis
C88-2123 there are many cases where we need feature percolation by f-rules . ( ii ) I will make
P90-1025 complex-feature based grammar in which feature percolation and agreement are distinct .
P90-1025 in these cases . The first is feature percolation and the second is the use of
E93-1034 taken as strong motivation for feature percolation ( see -LSB- Pollard , 1988 -RSB-
P06-1023 negative effects on the slash feature percolation mechanism . The probabilities
P06-1023 PCFG parser which applies a slash feature percolation mechanism to generate parse trees
J88-2012 frameworks . Some aspects , like feature percolation and control in GPSGs , are probably
J88-2011 frameworks . Some aspects , like feature percolation and control in GPSGs , are probably
P90-1025 if it is valid , the theory of feature percolation needs to be modified to allow
E99-1029 the subject without the need for feature percolation . On the other hand , in the
E99-1029 ) with re-entrant features and feature percolation , respectively . Constituents
E99-1029 potential to eliminate some or all of feature percolation , and in the remainder of this
E93-1003 way as some formalisms permit feature percolation principles to be separately stated
E93-1034 gapping . It is entirely unclear how feature percolation could engage such a construction
hide detail