Sound-to-print consistency of picture labels was manipulated in five experi
ments to investigate whether phonological codes constrain the selection of
orthographic codes in written picture naming. In Experiments 1 and 2, parti
cipants wrote down picture names which were inconsistent or consistent in t
he phono-orthographic mapping defined either at the level of the word unit,
i.e., heterographic homophones versus nonhomophones (Experiment 1), or at
the sublexical level (Experiment 2). In neither experiment did phonographic
consistency affect written latencies. Although more errors were observed f
or inconsistent than for consistent picture names, the observation of a sim
ilar error pattern in an untimed written picture naming (control) task sugg
ests that errors resulted from inaccurate orthographic knowledge. In Experi
ment 3, the position of the inconsistent units within the picture name (ini
tial versus middle or final) was manipulated. The results indicated that on
ly initial inconsistencies affected written latencies. Ruling out the hypot
hesis that this finding merely results from the fact that handwriting start
s before the orthographic encoding of the word endings, Experiments 4 and 5
showed that middle or Final inconsistencies influenced written latencies i
n a spelling-to-dictation task. The findings are discussed as suggesting th
at the build-up of orthographic activation from pictures is phonologically
constrained through the sequential operation of sublexical conversion. (C)
2001 Academic Press.