Speech perception is the process by which we convert a stream of speech into individual words and sentences. The objective when either listening or reading is to understand what is being communicated. Prosody refers to the rhythm, intonation and stress patterns in speech.

There are few clear boundaries between words in spontaneous speech and sounds blend together as they are produced. Words in speech are not presented as distinct units. We understand everyone’s speech as language, although there are a lot of differences between people, such as sex, age, accent and so on. Recognition of the word precedes the completion of the heard word. Research into the amplitudes of words shows that there is no clear boundary between words.

There are two major problems of speech perception:

1. Invariance problem
   This problem is hearing the same sound, although the physical properties of the sound are different (e.g: the physical properties of the sound are different when a 40-year old man speaks than when a 12-year old girl speaks, yet we recognize the phoneme as the same). The phonemes don’t change for us, although the physical properties do change. Co-articulation is the tendency for a speech sound to be influenced by sounds preceding it or following it. This will affect the following sound, yet we recognize the phoneme as the same.
2. Segmentation problem
   This problem is detecting distinct words in a continuous string of speech sounds, because there are no clear boundaries between words, although we do perceive those boundaries. The context is very important for recognizing words. The stress pattern of the language can be an important cue for recognizing word boundaries.

The foreign accent syndrome is a syndrome resulting from brain injury in which a person’s speech sounds like the accent of a foreigner. Infants tend to show a preference for their native language over an unfamiliar language. The development of word recognition requires the extraction of the regularities in a language that can be reliably used to distinguish word boundaries. Phonotactic constraints describe the language-specific sound groupings that occur in a language (e.g: some things are not allowed in a language: no English word starts with ‘XR’ and this can give a cue about word boundaries). Onset of a word is the initial phoneme or phonemes.

Slips of the ear occur when we misperceive a word or phrase in speech. It occurs when there is a misperception of a word boundary. These kinds of slips are also called mondegreens. Segmentation of incoming speech is biased towards the dominant patterns of the native language. Disfluency in speech can aid comprehension, as it can give the listener a cue that a less predictable word is coming.

Categorical perception is categorizing the incoming sound in a known category. This helps counteract the invariance problem. We are more sensitive to differences in speech sounds across phonetic categories than within. Voicing is when speech sounds are produced while the vocal cords are vibrating. People will categorize incoming sounds, even though it is rather different from the sounds they’re used to and they categorize it into categories of known phonemes. Infants can distinguish between the speech sounds of many languages at a young age, but this ability disappears as they acquire experience of the sounds of their own native language.

The left hemisphere language areas are accessed more efficiently by stimuli presented to the right ear. The right ear advantage for speech sounds refers to the finding that language sounds are processed more efficiently when presented to the right ear compared to the left. The effect of context can lead to the perception of absent speech sounds. The phoneme restoration effect describes the tendency to hear a complete word even when a phoneme is missing. This is a top-down influence, as the context drives the perception. Visual cues are also important for speech perception. The McGurk effect is a perceptual illusion that illustrates the interplay of visual and auditory processing in speech perception (e.g: if we see someone’s mouth make the movement for different phonemes than we actually hear, we are more likely to hear the phonemes that fit the movement of the mouth).

There are multiple models of speech perception:

1. Cohort model (autocorrect model)
   This model assumes that incoming speech sounds have direct and parallel access to the store of words in the mental lexicon. Some words can be recognized based on partial information. We establish expectations regarding likely target words once we have heard the initial phonemes of a spoken word. As more phonemes follow as input, fewer words remain and when there is only one option left, the word is clear to the listener. This is the uniqueness point. Context only plays a role at the moment that the sentence meaning is determined.
2. TRACE model
   This model assumes that top-down processes play a key role in speech perception. This model has three layers: features, phonemes and words. Features activate phonemes, phonemes activate words. Phonemes inhibit words that they’re not part of and words inhibit each other (e.g: there can only be one word at a time). The three levels of units follow a localist representation. A single unit represents a particular concept.

Lexical access is the process by which we access stored knowledge about words. There are two tasks that investigate lexical access. Word naming tasks require participants to name a word, while response time is measured and sentence verification tasks present a sentence frame with a target word and the participant must decide if the word fits in the frame. There are several effects that influence lexical access:

1. Frequency effects
   The frequency with which a word is used in a language affects cognitive processing, the higher the frequency, the easier the word is to process. These effects apply to open-class words, which are content words, but this does not apply to closed-class words, which are conjunctions, prepositions and these remain stable over time. People tend to fixate their eyes on low-frequency words longer.
2. Priming effects
   The task context can influence performance. If semantically related words are used in a lexical decision task, response time decreases. This is called semantic priming. Another type of priming effect is repetition priming Repeated exposure to a word leads to a faster response in a lexical decision task.
3. Syntactic context
   The syntactic category of the word and sentence context also affect lexical decision time. Participants are significantly faster in recognizing words when they occurred in sentences that provided the appropriate grammatical context.
4. Lexical ambiguity
   Homographs are words with the same spelling, but more than one meaning and pronunciation. When an ambiguous word is encountered, more than one meaning is initially activated and this affects the lexical access. Context does not affect initial access to multiple meanings.

The language used affects our perception of the situation (e.g: empty gas cans or gas cans). Parsing is the process by which we assign a syntactic structure to a sentence. We show a preference for one structure and interpretation and only when we realize that we made a mistake is when we go back. Reversible passives require more processing time. People with Broca’s aphasia don’t understand reversible passives. A phrase structure tree is a graphic representation of the syntactic structure of a sentence. Garden path sentences are grammatically correct but ambiguous sentences that bias the reader’s initial parsing. The goal of parsing is to assign incoming words to the appropriate role in the sentence as simply and efficiently as possible. There are two key strategies for this:

1. Minimal attachment
   This introduces new items into the phrase structure using as few syntactic nodes as possible.
2. Late closure
   This attaches incoming material to the phrase that is currently being processed.

Garden path sentences require the person to revise their initial interpretation of the sentence, as new, conflicting, information is presented.

There are four main types of scripts:

1. Logographic (ideographic) scripts
   These scripts represent morphemes or the units of meaning of words.
2. Syllabic scripts
   These scripts use a symbol to represent each syllable.
3. Consonantal scripts
   These scripts represent the consonants of the language.
4. Alphabetic scripts
   These scripts represent the phonemes or sounds in a language.

The scripts can also be combined. The grapheme is the written representation of a phoneme. Transparent or shallow orthography uses a one-to-one correspondence between the letters and sounds. Opaque or orthographically deep languages are those where the relationship between letters and sound is more complex.

The word superiority effect refers to the finding that a target letter within a letter string is detected more readily when the string forms a word. We can fill in letters if they’re not there when reading. Context has a considerable influence on visual word recognition. People cannot read a word and when asked to name the colour of the word, instead of the word, interference often arises. This is called the Stroop interference. Saccades are fast movements of the eye made when reading or scanning an image. Fixation time on a word is reduced if the reader has managed to preview the word prior to fixating it and fixation time is reduced for words that are readily identified. As the text difficulty increases, the saccade length decreases and the number of regressions increases. The processing cost is greater for phonological textisms compared to orthographic textisms.

Coltheart’s dual-route model of reading proposes three routes for reading:

1. Grapheme-to-phoneme conversion route
   You go from a letter to a speech sound and this is combined in a speech signal. This is necessary for the reading of unfamiliar or new words and can be used to sound out regular words (words that follow the print-sound rules).
2. Lexical route
   Reading happens via word recognition. It makes use of the mental lexicon. The sound properties are accessed because the word is stored in the mental lexicon.
3. Non-semantic reading route
   This route accounts for occasions when an irregular word is read correctly when the meaning is not available.

Pure word deafness (compromised route 1) is a deficit affecting the ability to recognize speech sounds, while comprehension of non-speech sounds remains intact. They can hear other sounds, just not speech. Other aspects of aphasia are absent. Pure word meaning deafness (compromised route 2) is a disorder in which the patient can repeat back the word, but cannot understand it.

Acquired dyslexia (alexia) refers to reading difficulties following brain injury. Surface dyslexia (route 1 is used) is characterized by a deficit in the reading of irregular words, while the reading of regular words is spared. Phonological dyslexia affects non-word reading, but real words can be read (route 2 is used). People with non-semantic reading can read regular and irregular words, but there is no comprehension of the words.