How does recognizing, understanding and speaking of language work? - Chapter 10

For cognitive psychologists, one of the most intriguing human abilities is the ability to use language. Steven Pinker (1994) described this as follows:  'Simply by making noises with our mouths, we can reliably cause precise new combinations of ideas to arise in each other's minds'. The question that arises is which cognitive processes are involved in the ability to use language. Many cognitive psychologists state that language skills cannot simply be understood in terms of memory, reasoning, and other cognitive processes; they argue that language ability is based on a relatively autonomous set of capabilities that function primarily independently of other cognitive processes. How the language system works has become especially clear by studying patients with language disorders. Some important concepts when it comes to language are:

• Speech is the spoken form of language. It is a way to convey linguistic information with the human voice;

• Syntax is about the grammatical rules of a language. These rules determine the way words can be combined. It is independent of the meaning of the sentence;

• Writing is a visual system to display the language. There are various writing systems, such as alphabetical, syllabic or ideographic / logographic.

What does the language system consist of?

A question that concerns psychologists is which processes are involved in:

• Understanding language (speech and reading);

• The production of language (talking and writing).

Linguistics has provided insights into the different levels of language, and the systematic and rule-led way it works. The following levels can be distinguished):

• Phonetics - the sounds of speech.

• Phonology - the sound system of a certain language.

• Morphology - word formation.

• Syntax - the combination of words into sentences.

• Semantics - the meaning of words and sentences.

• Reason (discourse) - which goes beyond individual sentences (stories, conversations, etc.).

Speech sounds

The field of phonetics studies the process of articulation of speech and the physical properties of speech such as sound waves. A phoneme is a term that expresses the smallest unit of sound that can make a difference in meaning. Phonemes can be combined to form words; every language has its own rules for this.

Visual language

Sign language is a visual language that mostly people who are deaf use. In sign language, the hands are used to express linguistic information. The face takes on the role of prosody and intonation in spoken language. It ensures the transfer of emphasis and emotion. Visual sign language consists of different shapes and movements of the hands and the fingers.

Words

Words (also referred to as lexical terms) are names for things, people, abstract concepts, events and properties of objects. So-called morphemes can be distinguished within words. A morpheme is a part of a word with its own meaning that cannot be broken down into smaller parts. A distinction can be made between free and bound morphemes. A free morpheme also occurs as an independent word; a bound morpheme occurs exclusively in combination with one or more free morphemes. Bound morphemes that are placed in front of a free-morpheme, are called prefixes. Bound morphemes that are placed behind a free morpheme are called suffixes.

Within free morphemes a distinction can be made between:

• Grammatical function words such as "the" and "and". This is a fixed set of words;

• Content words that express the semantic content. New words can be added to this group, so it changes over time.

The mental lexicon finally, the mental store of words; every word has lexical access and a specification of its meaning, pronunciation and function in the sentence.

Sentences

Words are combined into sentences. The syntax (sentence structure) contains rules for combining words in the correct order and for adding inflections at the end of a word. Phrases do not stand on their own. A sentence connects to a preceding one and together they build up a treatise. Language is a social activity that enables us to communicate and share thoughts with other people.

What does the study of psychology and linguistics entail?

The psychological study of language is often referred to as psycholinguistics. The scientist Chomsky has provided much insight into the complex language system. An important insight from him was that people routinely produce and understand utterances that are completely new to them. This creativity implies that language users have a set of rules. This set of rules is referred to as grammar.

Chomsky also stated that each sentence had two grammatical classes: a 'surface structure' and an 'underlying structure'. The 'surface structure' is about the form in words; the 'underlying structure' indicates the meaning. Two sentences can be very different in 'surface structure', but very similar in 'underlying structure', and vice versa. A question that can be asked is whether there is a relationship between the syntactic complexity of a sentence and the effort it takes to process a sentence. Research indicates that it takes more time to process complex sentences than simpler sentences; however, this only applies to phrases that are reversible (in the sense that subject and direct object can be exchanged, without creating an incomprehensible sentence). Non-reversible sentences can be processed faster than sentences that are reversible.

Language behavioral research requires an understanding of the underlying cognitive and linguistic processes. This can be done through research with tasks, the lexical decision task is one of them. In this experiment the participants are given a (mostly written) target item, then it must be decided whether it is an existing word or not. These types of tasks are used to examine the amount of time required to make the decision. This information provides an indication of how the item is being processed.

How do we recognize spoken and written words?

To understand what people are saying, the ability to recognize individual words of a message is crucial, and this process is very fast. There is evidence that there are separate systems for spoken and written word recognition. Cognitive psychologists are particularly interested in the question of which processes make such efficient word recognition possible. The main factors that influence the efficiency of word recognition are:

• Word frequency effect: High frequency words are recognized faster than low frequency words, these are words that are used a lot;

• Context effect: The term 'semantic priming' is related to the context. Semantic priming is used to determine whether the meaning of a previous word or sentence influences the recognition of a subsequent word. Research shows that 'semantic priming', and therefore the context, is of influence.

• Neighbour effect: How many words can you make that are different in just one character from the original word. The more neighbours, the faster the word becomes recognizable.

The previously mentioned term 'mental lexicon' plays a central role in explaining the process of word recognition.

Listeners and readers approach the lexicon every time they hear or read a word. The question that arises is how we can get access to the relevant word in the lexicon so quickly. Two types of models have been drawn up for this:

• Direct access models: these models assume that each item from the lexicon can be approached separately and can be looked up directly. An influential model of this type is Morton's so-called 'logogenic model'. Word detectors (logogens) monitor evidence for both the stimulus and context (this is a passive process); recognition occurs when the proof of a certain word exceeds the activation threshold. Another well-known model is the 'cohort model'. This model was developed in an attempt to explain the process of spoken word recognition. It assumes a sequential nature of word recognition: Based on the initial sound, a set of possible words is activated,subsequently more and more items are removed from this set as more sounds of the input become available (this is an active process).

• Serial search models: these models assume that the items in the lexicon are passed through one by one until it matches the searched input.

Spelling is an important part when it comes to written words. Regular spelling ('regular orthography') is a writing system in which there is a direct correspondence between speech sounds and letters. In the case of irregular spelling ('irregular orthography') there are less consistent spelling to sound rules. This means that the relationship between speech sounds and letters is more variable and more opaque.

The classical approach in cognitive psychology is to keep control over as many constant factors as possible and then to manipulate a factor in isolation. This is easier said than done when it comes to the complex processes that come up during word recognition. Researchers can also unknowingly influence the list of words (bias). Finally, it has emerged that only multisyllabic words are used during studies.

How do we understand words?

There are many theories about how humans understand the meaning of words. In a modular system, it is stated that semantic information is encoded in abstract representations. It is not yet clear what these representations consist of: are they prototypes, hierarchies of information or examples? A theory states that semantic knowledge is encapsulated, there is a more distributed quality of semantic representations. The semantic representations of words can be associated with sensory motor information. The word glove, for example, can be presented visually and emotionally (how it feels to put on a glove). Concrete nouns have more sensory associations.

What is lexical access in language comprehension?

Several factors apply to word recognition and there are a number of models that represent our mental lexicon (all words we know). In the John Morton's Logogen Model, each word consists of a unit, which is also called a "logogen." Each log represents representations of a word. When the perceptual systems process incoming words, the activity of logging increases. Contextual evidence can also start this activation, because if a word starts with 't' and this is said in a furniture store, the word 'table' will be activated earlier than 'title' even though both words start with 't'.

An alternative model is the cohort model of Marslen and Wilson that has focused on the recognition of spoken words. The model describes how visual and auditory information is stored in a person's lexicon. When a person hears words, each word will activate segments in the lexicon that begin with that segment. If multiple segments are added, more words are deleted until the word that matches the information is found.

Many of the cognitive models in this summary either belong to the 'connectionist', 'parallel-distributed models' or 'rule-based systems'. This is the case for cognitive - linguistic models, for example models about the concept of 'reading' can be connectionistic or symbolic.

How do we understand sentences?

Understanding language involves more than just recognizing words. Syntactic and semantic knowledge are needed to determine what certain words mean in a certain context. It is also important to find out how these meanings can be combined. In addition, knowledge of the world around us is sometimes needed to grasp the meaning. It is difficult to investigate language comprehension because the processes are almost automatic. Phrases are often already fully understood, even though the sentence is still incomplete. To gain insight into the processes involved in the interpretation of sentences, psycho-linguists have made considerable use of sentences that have an ambiguous meaning. With regard to this ambiguity, a distinction can be made between:

• Structural ambiguity: this is the case if the sentence as a whole leads to several possible interpretations;

• Lexical ambiguity (lexical ambiguity): this is when a sentence contains an ambiguous word. A word can therefore have several meanings.

In this context the so-called garden-path sentence emerges. An garden-path sentence is an intentionally misleading sentence that plays with metaphorical meanings of words or with homonymy (the same word that can have two different meanings), so that a reinterpretation occurs during reading. An example of a garden-path sentence is: "I saw people with those big binoculars." A question that can be asked is how we process sentences that we hear and see:

• ' Word-by-word', also known as the 'immediacy principle' (direct processing word by word);

• ' Wait and see' (wait until a sentence has been completed before processing it).

Because the second approach places great demands on the working memory, it is most likely that the processing takes place through the 'word-by-word' approach. Decisions about how each word functions are then made gradually. There are two points of view on how syntactic, semantic and pragmatic information is applied in the interpretation of sentences. The first point of view is that the syntactic analysis of the sentence is independent of, and not influenced by, semantic and pragmatic knowledge.

The second position is based on interaction; this states that semantic information can guide syntactic analysis.

Frazier believes that syntactic analysis is autonomous and that it operates independently of semantic and pragmatic issues. However, this opinion is controversial; other psycholinguists argue that there is an interaction between the process of syntactic analysis and information processing based on semantic and pragmatic knowledge. Much research into the processing of sentences is based on ambiguous sentences or so-called garden-path sentences, presented in isolation outside any context. More recent research, however, also focuses on more general and contextual sentences. Grodner (2005) did research into the influence of the 'discourse' context on sentence comprehension. His research showed that the 'discourse' context influences the syntactic process; according to him, there is interaction between processes.

How do we produce language?

When people talk, their thoughts are turned into language. Levelt (1989) states that lexicalization means the process by which the thought underlying a word is converted into the sound of the word. He distinguishes two phases in lexicalization:

1. Lemma phase: construction of the grammatical form and the semantic representation;

2. Lexeme phase: specification of the phonological form.

The selected 'lemma' activates the 'lexeme'. The state of 'something is on the tip of my tongue' (tip-of-the-tongue phenomenon) can be seen as a successful 'lemma' phase, but as a failed activation of the 'lexeme' phase. In academic circles, the question is whether the two phases are independent of each other or whether there is some kind of interaction. Language production has mainly been studied on the base of lapsus (slips-of-the-tongue). In practice, discussions appear to be fairly systematic. They usually relate to one word, one morpheme or one phoneme, which is substituted for another word, morpheme or phoneme. Garrett divided the discussions into four groups:

• Word substitutions ('boy' for 'girl', 'white' for 'black', etc.). Only occur with content words;

• Word changes ('I write a mother to my letter' instead of 'I write a letter to my mother'). Appear in words from the same category;

• Sound changes - sound exchange errors ('big cave' instead of 'big god');

• Morpheme changes - stranding errors ('a maniac for weekends' instead of 'a weekend for maniacs').

Garrett drew up a model of speech production. According to him, a number of different levels are involved in producing a sentence that operate independently of each other. He distinguishes the following levels from top to bottom: 'message level', 'functional level', 'positional level', 'phonetic level' and 'articulation level'. The message level includes the concept you want to talk about, but this is not defined in language, but in places. At the functional level, concepts get an agent and an object and are linked to semantic representations. In the positional level, words are placed in the correct order. Garrett assumes that speech production is a serial issue in which the different levels operate independently of each other and run from higher to lower. Other researchers have their doubts about this idea and state that there is an interaction between the different levels. For example, Dell states that during the speech production process, scheduling takes place simultaneously at semantic, syntactic, morphological, and phonological levels. According to him, the processes are therefore not independent and not serial. Syntactic, morphological and phonological levels planning takes place. According to him, the processes are therefore not independent and not serial. 

Levelt came up with the six stage model of spoken word production, called WEAVER ++. The six stages of this model are as follows: conceptual preparation, lexical selection (the abstract lemma is selected), morphological encoding, phonetic encoding and articulation. The model is serial, which means that one stage must be completed first before heading to the next stage.

What is the 'discourse' level?

The term 'discourse' is used to indicate that a number of sentences are related to each other in a meaningful way. When studying language at the 'discourse' level, we look at the way in which speakers and listeners:

• integrate the thoughts and ideas expressed in sentences into topics;

• switch from one topic to another.

There are some universal rules that are followed when we have a conversation with someone:

• Change of speaker takes place;

• Only one speaker speaks at a time;

• Simultaneous speech is short-lived in use;

• Transitions without a gap (for example, silence) and without overlap are common;

• The turn order of speaking is not fixed, nor does it have a fixed duration, length or content;

• The number of speakers can vary;

• The conversation can be continuous or discontinuous.

Psychologists distinguish between given and new information. Information provided refers to information that the speaker assumes the listener already knows; of new information, the speaker assumes that the listener does not yet know. Information given is often introduced by the definite article 'the'; new information through the indefinite article 'a'.

For smooth communication, the speaker and the listener must try to understand each other:

• the speaker will have to estimate what the listener already knows about the topic and what needs to be further explained;

• the listener will make an estimate of where the speaker wants to go and why something is being told.

Grice (1975) stated that it is important to also understand the intention rather than just the speaker's words to understand him or her well. He called the various implied meanings that he distinguishes 'implicatures':

• Conventional implications : these arise from semantic representations of the meaning of certain twists;

• Conversational implications : these stem from the concept that there is implied meaning in how sentences are constructed in speech;

• Cooperative principle: this means that you have to make a conversational contribution as needed, at the stage that takes place, by the accepted goal of the direction of the conversation you are currently in.

Language not only has the goal of sharing information with others, but also has the goal of making and maintaining social connections. Social conversation is the name for this last goal. Language allows us to share information and to have social interactions. We are social primates, and we want to show ourselves and others what our most important social connections are. Direct (meeting up) or indirect conversations (through a computer) take longer than conversations by telephone or by e-mail.

For cognitive psychologists, one of the most intriguing human abilities is the ability to use language. Steven Pinker (1994) described this as follows:  'Simply by making noises with our mouths, we can reliably cause precise new combinations of ideas to arise in each other's minds'. The question that arises is which cognitive processes are involved in the ability to use language. Many cognitive psychologists state that language skills cannot simply be understood in terms of memory, reasoning, and other cognitive processes; they argue that language ability is based on a relatively autonomous set of capabilities that function primarily independently of other cognitive processes. How the language system works has become especially clear by studying patients with language disorders.