Similarly to lahguage aquisition, language processing faces a strong input-data-deficiency problem. When we speak we alter a great deal in the idealised phonetic and phonological representations. We delete whole phonemes, we radically change allophones, we shift stresses, we break up intonational patterns, we insert pauses in the most unexpected places, etc. If to such crippled phonological strings we add all background noise which does not help comprehension either, it is difficult to imagine how the parser is supposed to recognise anything at all. However, even in the most difficult circumstances (foreign accent, loud environment, drunkenness, etc.) we do comprehend speech quickly and efficiently. There must be then some signals in the phonetic string which are particularly easy to grasp and to process. We call these signals PIVOTS and parsers working with these signals we call PIVOT PARSERS.