Schemas or schema theory
In TESOL or teacher training courses, the term schema is often described or presented merely as background knowledge, e.g., the students' background knowledge of a particular topic. So one hears of schema activation, a sort of introduction, segue, or hook at the beginning of a lesson to activate relevant background knowledge of the topic or theme of the lesson. This is a very deficient understanding of schemas, as the term is properly understood in psychology. All information is encoded and stored as schemas in the human mind. A schema is not just background knowledge, but all knowledge that is organized.
A schema is defined as an organized mental concept, an organized mental representation for concepts and information. All our concepts and knowledge stored in our minds is organized, both internally and externally. It is internally organized, in that it consists of structured information on the concept, such as its properties, components, features and such. It is externally organized, in that it is connected with schemas for related information, associations, categories, and hierarchies of concepts. For example, every word and concept in our minds is schematic, from large to small, from concrete to abstract, e..g, what we know about birds, dogs, love, bridges, winter, feminism, happiness, one' childhood, why we have seasons, the earth being round, gravity, how one gets food in a restaurant, what a father is, other people, one's one sense of self and self-esteem, and whether you think you can successfully learn Latin.
Everything that we know is stored in our minds in the form of mental concepts. These mental concepts are highly structured and interconnected, and are known as schemas. There are different kinds of schemas for different types of concepts or different types of conceptual knowledge. All our knowledge is schematic. The term ‘schema' comes from the Greek word for framework (plural: schemas in American English, schemata in older or more British contexts), as a mental concept consists of a number of details and pieces of information, with some being more peripheral and secondary, and everything being organized around more core or central information.
For any word and concept such as dog or love or tall or anything else, we are dealing with both the word and the mental concept, which are both closely connected. The conceptual aspects of these schemas in turn consist of different kinds of schematic information. Theorists talk about different types of schemas, which is a useful distinction. But we also have to proceed with caution, because it is not always possible to make clear-cut distinctions or divisions.
For example, your general schema for ‘dog' consists of the conceptual concept and the linguistic content (lexical schema). The conceptual content (or semantic content) consists of exemplar and prototype information, and probably other content as well. These different types of contents are closely interconnected, and it may not always be possible to distinguish them.
The following are characteristics of schemas.
- Structure. A schema consists of information that is organized and interconnected.
- Internally organized. A schema can consist of different types of information. It can also consist of subschemas or subtypes. For example: the dog schema includes different types of dogs (hounds, Great Danes, chihuahuas, German shepherds, etc.)
- Externally organized. A schema is conceptually connected to related schemas (e.g., dog → cat, pets, animals, mammals, wolves, coyotes...), including category information (table → furniture).
- Complexity. A schema consists of complex and highly organized conceptual knowledge, consisting of different types of conceptual information (e.g., ‘dog' → prototype, exemplar, functional information), which are all interconnected.
- Categorization. Schemas can belong to hierarchies with associated concepts, superschemas, subschemas. These allow us to categorize concepts and things that we see in the world. Nouns, for example, can be lower members (hyponyms) of larger classes (hypernyms, or super-ordinate level terms); e.g., the hypernym ‘furniture' includes ‘table, chair, bookshelf...'
- Associations. A schema consists of associations of information, not just list or pile of information or definitions (as in traditional semantics); it consists of a complex set of features and other information. For example, for ‘dog,' all the prototype features and information, exemplar information, memories, linguistic (lexical) content, and other pieces of information, are all interconnected in the mind, and in the brain, in the various neurons that encode all this information in the brain.
- Experiential. Schemas are formed from our experiences and the input into our minds. Many things are stable and shared across individuals, to the degree that we have similar brains experiencing the same world, so for most concepts, we have similar shared knowledge. We all encounter the same sun, water, and sensation of the color red. However, some aspects are molded by our culture, which affects our experiences and how we filter them. So we might have slightly different exemplar schemas for dogs (as small dogs are more normative in some places and larger dogs in others), or we might name basic colors in the blue-green color spectrum differently. Schemas, then, depend on our world and culture.
- Extensible. Because of all the associations in a schema, this makes it possible to learn new things by activating relevant features of an existing schema to create a new one and two connect new information with what we already know. We can also use this in new situations, by activating relevant features of an existing schema that are similar to a new situation, so we can make predictions about what to do.
- Emergence: Whole > Σ parts, or the whole is greater than the sum of the parts. With man-made devices, the whole is equal to the sum of its parts (whole= Σ parts). For example, we can assemble a computer, knowing what each component does, and knowing what the whole machine will do; what the whole machine does is predictable based on its components. For the mind and brain, it is different. For schemas, meaning arises (emerges) from all the bits of information being networked together in mental space, in ways that we do not fully understand. For the whole brain, consciousness and other things that make us human emerge from billions of neurons and trillions of connections, which we do not understand very well.
- Cognitively adaptive. All these features give human minds great cognitive abilities and computing power, allowing us to adapt to new situations and learn. They also give language a great deal of flexibility and expressiveness.
Cognitive adaptiveness is realized in the following ways.
- Prediction: E.g., guessing how to handle new situations; guessing the meaning of new words
- Learning new things, by making associations with prior knowledge
- Categorization: Classifying things in the world
- Interacting with and understanding our world
- Polysemy: Giving a word new meanings to handle new concepts (e.g., derived meanings of cup, table, dog, ship, cold...)
- Creating metaphors
- Expressiveness in language
- Grammar: Probably all grammar forms, structures and properties are schematic
2 General types of schemas
Schemas, or the information contained in a schema, may be of the following types. Some of these may overlap, or may be hard to distinguish or tease apart when we examine actual mental concepts. That is, some or many of our concepts may involve more than one type of these schematic types or structures.
- Prototype: Core features and/or abstracted mental image. An abstract representation, based on features and experiences, from which you've formed a generalized, core representation; e.g., your concept of bird is in part a prototype schema, an abstract representation of a typical bird, its features, and behaviors
- Exemplar: Key examples, including mental images, experiences, episodic memories, or sensory information. A particular example or collections of examples, e.g., your concept of ‘dog' is defined by typical dogs that you've known.
- Theory: An explanatory concept that is designed to explain something about the world. This can include common, everyday things (e.g., how the earth is round, object permanence, object support & containment, conservation of mass), as well as scientific and academic theories (e.g., gravitational theory, theory of evolution).
- Frames / scripts (procedural): Conventionalized ideas of how certain events are to be carried out, e.g., how you enter a sit-down restaurant and order food. A concept for how something is performed or done; e.g., what you do in a restaurant to get food; what you do in a living room; how certain kinds of conversations and social interactions proceed.
- Event schemas: How particular events and episodes happend[ed], such as your memories of your past experiences, along with your interpretations of those events.
- Relational. Some concepts are defined in relation to something else; e.g., father → children; tall → short. Some terms are defined in relation to others, e.g., ‘father' can only be defined in relation to ‘children'.
- Functional: What something is used for; e.g., one's schema for ‘hammer' consists of what it does and what it is used for. Some terms / concepts are defined in part by what they do, e.g., specific devices, service personnel.
- Image: An image schema is a representation of the physical characteristics of a concept. This would not only include physical objects and their physical properties, but more abstract properties. For example, we associate "up" with an increase in amount (e.g., because containers become fuller and contain more as the amount physically goes up in the container), and by extension, with less concrete types of increase. We can thus extend this to ‘up' in a more qualitative sense, like prices going up.
- Lexical. Words and word meanings. The linguistic content connected to a concept, e.g., words like dog, table, love, peace, die. This would include grammatical information (word class – noun, verb, or such; what kind of noun or verb it is; how it is typically used in sentences; other words it is used with; its frequency in the language; as a noun, whether it is commonly used as a sentence subject or object, and with what kinds of verbs; as a verb, the types of nouns that it takes as subjects and objects); morphological information (word parts); phonology; nuances and usage in context; and the associated semantic content.
- Grammatical. Grammar structures and features are schematic. This includes specialized grammatical schemas, such as argument structure configurations.
- Social. Personal, social psychological, and related structures (see below).
- Discourse. While reading or listening to discourse, we construct an online, ad hoc schema, which is updated as we read or hear more. We draw from existing schemas to build a mental representation for the discourse (conversation, lecture, etc.) that we are currently reading or hearing.
The idea that grammar is essentially a set of specialized schemas forms the basis of Cognitive Grammar and Construction Grammar.
3 Discourse schemas
Schema are important not just in interpreting information, but also in decoding how that information is presented. Schemata can be reflected in text structures, for example, (Driscoll, 1997; Halliday & Hassan, 1989). Readers use their schematic representations of text (narrative, compare/contrast, cause/effect, etc.) to help them interpret the information in the text. Schema reflecting how information is presented can also be culturally determined. Kaplan (1966) stated that the structure of formal argumentative essays is culturally determined and that therefore second language writers and readers must be aware not only have sufficient command of their second language but also of the textual structures in their second language.
Discourse comprehension involves several levels, with the preliminary stage being the initial input stage before work on the schema level begins. The eyes and mind take in information by focusing on key words, mainly content words (nouns, adjectives, adverbs, verbs), and devoting attentional resources to identifying the words – usually 200-400ms per word in reading.
The first sentence is just one word, hence, one concept or "proposition" that [X is cute], but the reader has to activate <cute> in the mental lexicon and figure out how it relates with what came before it, which includes inferring how the character meant it sarcastically. After integrating it with the reader's conceptualization of the discourse, the reader's working memory is cleared and is ready to process the next sentence. The next sentence activates concepts like [how come, ignored => (you) ignored (us)], which are buffered until the end of the sentence, then integrated with the reader's understanding of what has happened, e.g., connecting this question with how B avoided A and his friends before. In the next sentence, after identifying the new concepts, the reader connects it with the preceding discourse and events of the novel, and makes inferences, e.g., that this sentence is a defiant response to A's question.
One's discourse model for a reading passage or conversation consists of a structured mental representation for components such as:
- Topics: the topic / subject of the current sentence, topics of recent sentences, recently mentioned new concepts that are also important to the current sentence, and the global topic of the whole discourse.
- Characters: Characters in the story, or participants in the conversation, and their roles, actions, intentions, etc. (e.g., as good guy or bad guy in a story), including various assumptions or inferences you make about them
- Events: Sequences of events in the story, and how the relate to each other and to the characters / participants
- Context: Where and when the story takes place
After sentence final integration / wrap-up, syntactic info and verbatim memory of the exact words of the sentence are generally discarded; new information is integrated into online the mental representation. Processing of sentences & discourse is affected by age, cognitive impairments, and second language ability.
Theories of discourse schemas in psycholinguistic and psychological research are known as situation models or construction-integration models, most notably, those from (1) Kintsch and van Dijk's construction-integration model, and (2) Gernsbacher.
4 Social schemas
Our concepts and understandings of our social world. These allow us to efficiently assess and understand social situations and people -
- Allow us to interpret others' behaviors, actions, & intentions
- Allow us to understand and interpret social interactions
- Allow us to communicate and interact with others
- (General) social schemas are about general social knowledge – e.g, how people generally are, how they interact; this includes assumptions like the "theory of mind" - our assumptions that other people are intelligent, social beings like ourselves, with intentions, feelings, conscious minds, etc.
- Person schemas are about individual people. These include schemas for particular individuals that you know; idealized, prototype person schemas for new people or strangers; schemas for personality traits; for types of people in different social contexts
- Stereotypes & social group schemas: Your concepts and assumptions about particular groups of people – e.g., your concepts of males cf. females (gender schemas); old people; foreigners; blacks; bankers; businessmen; professors; etc.
- Self-schemas: Your concept of your own self, which consists of different sub-schemas for different aspects of your identity, and for different social roles and situations.
- Role schemas: Concepts of proper behaviors or expected behavior in given situations.
- (Social) event schemas (or scripts): For what happens in specific social situations and interactions
Schemas can acts as "filters" thru which we interpret the world; e.g., in a social psychology study by C. E. Cohen (1981), subjects were shown a video of a librarian drinking, and later recalled & reconstructed the scene of the librarian drinking wine, as influenced by their views of typical librarians. People hold on to their schemas, even when faced with evidence to the contrary. Stereotypes are particularly powerful – e.g., people who don't like police officers will interpret any actions by a cop negatively. Motivations and expectations can also cloud our judgment & reasoning skills. Likewise, intellectual or emotional stimulation (arousal) toward a certain viewpoint, person, event, etc. can alter our thinking, for example, cognitive dissonance. Influence of stereotypes (Bargh, Chen, Burrows 1996): when primed for concepts of rudeness, subjects behave more aggressively with experimenters, or when primed for elderly stereotype, subjects walk more slowly after experiment.
5 Pedagogical implications
- No rote learning!
- Introduce new concepts, vocabulary, grammar patterns in meaningful contexts - story, conversation, etc.
- Give concrete / simplified examples & explanations
- Use mnemonics
- Allow time for review, repetition, etc.
- Have students ask questions
- Make students discover & practice on their own
- Learning in context is more efficient
- Let contextual schemas serve as aid to new vocabulary, concepts, structures
- Focus more on smaller, more meaningful set of words
- Use realistic examples; don't rely so much on artificial textbook examples
- Reading/writing & conversational practice required for developing associations
- Teach grammar as patterns, not rules – patterns that convey meanings, rather than mechanistic rules. Any grammar form or structure is a pattern that conveys a certain meaning, from more straightforward (verb tenses, determiners) to more abstract and complex concepts (e.g., the perfect tense, and the determiners a and the).
- Try to present new information in the form of patterns rather than just rules. When possible, present new concepts in the form of exemplars / models, prototypes, relationships, categories, events / activities, etc.
6 See also
- Beware, however, of exaggerated claims like strong forms of the Sapir-Whorf hypothesis, that claims that culture and language determine one another. Beware of claims related to this view, e.g, that Eskimos have dozens or hundreds of words for snow. They have one or two basic words for snow like we do; like us, they have other non-basic snow and weather-related terms like slush, sleet, blizzard, and such.
- The term ‘event schema' in other contexts can also refer to non-social schemas for types of events and actions we see in the physical world, like cause-effect events.