According to D.W. Massaro, author of Speech Perception (2001), this term refers to "the process of imposing a meaningful perceptual experience on an otherwise meaningless speech input." The simplest definition of speech perception explains the term as "the auditory perception and comprehension of speech." Speech perception is regarded as one of the most important processes in oral language communication. Study in this field is focused on the process by which acoustic speech signals, in other words sound waves, are received and comprehended. Speech perception is a subject of interest in cognitive psychology and psycholinguistics. Research on the topic is limited to non-invasive data gathering techniques. In fact scientists have found solutions to the problems of speech perception, without being able to understand its nature.
One of the easiest and most effective ways to explain and grasp the concept of speech perceptions is for a person to first examine its connection to acoustic signal, phonemes, lexicals, syntax and semantics. Experts have concluded that speech perception in not simply about the reception of acoustic signals and its direct translation into semantic understanding. Through the principle of recursive decomposition, analysis of speech perception can start with the consideration that a sentence is composed of grammatically arranged words, which in turn are composed of syllables. However, experts have found that the basic unit of speech perception is neither the syllable nor the phoneme.
The latter conclusion is based on the observation that a single phoneme can be pronounced differently depending on its combination with other phonemes. The different pronunciation of phonemes is referred to as coarticulation. The articulatory gestures that are required to produce certain words blend the phonemes, and thus it becomes impossible to distinguish them as separate units. As a result experts point out coarticulated phonemes as the basic unit of speech perception.
Experts on this topic have developed a number of theories in their effort to understand the nature of speech perception. One of the popular theories is the motor theory. According to this theory, the listener distinguishes phonemes because of the fact that he himself is a speaker. Before the listener has heard any syllable, he or she knows what articulatory gestures have to be made in order to produce it. The motor theory states that speech perception is an innate, species-specific or special ability of human beings. Meanwhile, the auditory theory of speech perception appeared as a result of the study of animal communication. It challenged the motor theory with the conclusion that speech perception as a process is not unique to humans. According to the auditory theory, speech perception does not occur thanks to any special knowledge about the nature of how speech is produced. Instead it is derived from general properties of the auditory system.
Franck Ramus, from the Institute of Cognitive Neuroscience & Department Psychology at the University College London and American evolutionary biologist Marc D Hauser conducted studies on speech perception in human newborns and cotton-top tamarin monkeys at the beginning of the 21st century. The results of the study supported the auditory theory of speech perception. The scientists showed that human babies and tamarins can recognize the difference between languages. Both newborns and monkeys were found to notice a change in the language spoken and in the speaker.
It is important to note that perception and understanding are influenced not only by the actual sound of the speech but also by a speaker's face and accompanying gestures. This fact further adds to the complexity of the human speech perception system. As a result, regardless of the number of facts and experiments that support or oppose the motor theory and the auditory theory, there are some effects that both fail to explain. Many other fields of study have examined the phenomenon of speech perception. These include psychophysics, neurophysiology, sensory perception, psycholinguistics, linguistics, artificial intelligence and sociolinguistics.
Artificial intelligence might appear to be out of place when talking about speech perception. However, the development of a computer system capable of natural language processing is one of the ambitions of computer scientists as engineers. To achieve that ambition, scientists need a complete understanding of the physics which make human speech perception. As long as synthetic language processing remains a goal, speech perception will be a relevant topic for computer science.