📘 Lecture 12 — Schemata and Discourse Processing
📖 Overview: This lecture explores how schemata—mental structures organizing our knowledge—influence how we comprehend and remember discourse. It examines different discourse genres, the psychological validity of story grammars, and strategies for identifying main points in texts, making it essential for understanding the cognitive processes behind reading comprehension.
🗂️ Topics Covered
The lecture covers schemata and their role in memory reconstruction, discourse genres including narrative and expository forms, narrative discourse processing through story grammars with cross-cultural evidence, inaccessibility of knowledge during comprehension, methods for identifying main points in texts, and strategies for building global text structures.
📝 Lecture Summary
Topic-067: Schemata
A schema (plural: schemata) is defined as a structure in semantic memory that specifies the general or expected arrangement of a body of information. This concept is associated with Bartlett's early work on story recall. Bartlett demonstrated that remembering is not a rote process but a reconstructive one where we retain the overall gist and reconstruct details from that impression. In his experiments, when college students were given stories inconsistent with their schemata, recall was distorted toward the schemata—they either didn't remember discrepant events or normalized them, altering details to fit existing schemata. Though initially unappreciated, Bartlett's ideas have gained significance as modern research focuses on characterizing schematic knowledge more precisely and determining how it's used during discourse comprehension.
🔑 Definition — Schema: A structure in semantic memory that specifies the general or expected arrangement of a body of information.
📌 Example: When British college students were given Eskimo folktales inconsistent with their schemata, they recalled poorly because they normalized the stories to fit their own cultural expectations.
💡 Why this matters: Understanding schemata explains why people remember information differently based on their prior knowledge and expectations.
Topic-068: Genres
Genres are types of discourse that have characteristic structures. They provide general expectations about how information in discourse will be arranged. Examples include lectures, sermons, opinion articles, and presidential speeches. News articles follow an inverted pyramid structure—most important points in headline and beginning, with less important details later—because editors delete from the end when space is limited. Psychology students learn the APA format for journal articles: abstract, introduction, method, results, discussion. Narrative discourse (stories) typically introduces characters and setting, then follows a protagonist with a goal, obstacles, and resolution. Different genres exist for detective stories, fairy tales, and romances. Expository discourse contrasts with narrative—its goal is conveying information about subject matter in organized, logical manner, as in textbooks and lectures.
🔑 Definition — Genre: A type of discourse that has a characteristic structure, providing expectations about information arrangement.
📌 Example: In detective stories, writers create interest in a crime, supply possible motives for suspects, and drop clues—readers can imagine many outcomes at the beginning, fewer in the middle, and predict the ending.
Topic-069: Narrative Discourse Processing
Story grammars formalize Bartlett's ideas—they are schemata in semantic memory identifying typical event arrangements in stories. A story grammar views narratives as consisting of a setting, one or more episodes, and an ending. Each episode has: an initiating event, an internal response from the protagonist, a goal, an attempt, and an outcome. Evidence supports the psychological validity of story grammars. Black and Bower showed episodes are recalled in an all-or-none fashion—the length of one episode doesn't influence recall of another. Glenn reported that episodic structure of recall is unaffected by episode length. Cross-cultural investigations by Mandler, Scribner, Cole, and DeForest found that story recall patterns vary across cultures—Bartlett's British students recalled Eskimo folktales poorly because their schemata didn't match.
🔑 Definition — Story grammar: A schema in semantic memory that identifies the typical or expected arrangement of events in a story.
📌 Example: In a story, an episode might begin with a character seeing a fire (initiating event), feeling afraid (internal response), deciding to escape (goal), running to the door (attempt), and reaching safety (outcome).
Topic-070: Inaccessibility of Knowledge
Sometimes we fail to activate appropriate knowledge during comprehension. One type of inaccessibility occurs when text is written so obscurely that we cannot bring our knowledge to bear—this severely impairs comprehension. However, it's also possible to comprehend a passage without activating relevant knowledge. Garrison Keillor's story about Uncle Lew's tale demonstrates this: despite hearing the story dozens of times, the listener was never sure the children would escape the burning house. This shows comprehension can occur without full knowledge activation.
📌 Example: In Uncle Lew's story, a house burned down and children ran barefoot into snow—though heard repeatedly, the listener was never certain all would get out, showing knowledge inaccessibility despite comprehension.
Topic-071: Identifying the Main Points
Determining main points can be difficult, especially for individuals with learning disabilities. Studies show difficulty stems from distracting details. Meyer, Brandt, and Bluth (1980) found that when key points are signaled explicitly, performance improves—signals helped poor comprehenders (those not sharing the author's schema) but didn't affect good comprehenders. Strategy training led to increased recall performance, more effective than signaling alone. Reder and Anderson (1980) found that eliminating details (as in Cliff Notes condensed versions) improved retention better than standard textbook versions. Similarly, Giora (1993) found analogies in text didn't facilitate comprehension and may impair recall—we comprehend best when extraneous material is omitted.
📌 Example: Reder and Anderson found that students remembered more from condensed versions of plays than from full textbooks—eliminating details improved retention.
Topic-072: Building Global Structures
Ultimately, we need to identify important points even without explicit markers. As we become familiar with an author's content and structure, we can deduce the author's schema. A good test is writing a summary—this requires selecting important propositions and generalizing individual propositions into broader thematic statements. By comparing our summary with the author's, we can gauge how close we've come to extracting the gist. As proficiency increases, we can rely more on global processing strategies.
📌 Example: Writing a summary of a textbook chapter forces selection of main points and generalization—comparing this to the chapter's actual structure reveals understanding of the author's schema.
⭐ Key Takeaways
Schemata are mental structures that guide comprehension and memory reconstruction—we normalize inconsistent information to fit existing schemata. Discourse genres provide expectations about information arrangement, with narrative discourse following story grammar structures (setting, episodes with initiating events, responses, goals, attempts, and outcomes). Knowledge inaccessibility can impair comprehension even when we understand individual sentences. Identifying main points is facilitated by explicit signals and eliminating distracting details, with strategy training being more effective than signaling alone. Building global structures requires deducing an author's schema, best tested by writing summaries that extract the gist.
🧠 Quick Revision Questions
- What did Bartlett's experiments with inconsistent stories reveal about how schemata function in memory recall?
- How does the inverted pyramid structure of news articles reflect the genre's characteristic organization?
- What are the components of an episode in a story grammar according to the lecture?
- What did Reder and Anderson (1980) discover about the effect of eliminating details from text on retention?
- Why did Mandler and colleagues find that cross-cultural story recall patterns differ between groups?
📘 Lecture 13 — The Psychology of Learning
📖 Overview: This lecture explores the nature of sign language as a true linguistic system, comparing it with spoken language to understand biological constraints on linguistic form. It covers the psychological reality of grammar, including transformational grammar and alternative lexical theories, examining how linguistic structures relate to cognitive processing.
🗂️ Topics Covered
The lecture covers sign language as a true language without speech, speech-based versus independent sign languages (ISLs), basic grammatical concepts including duality of patterning and morphology, insights from American Sign Language (ASL) including fingerspelling, transformational grammar evaluation criteria (observational, descriptive, and explanatory adequacy), and issues in grammatical theory including the derivational theory of complexity and the centrality of syntax in linguistic knowledge.
📝 Lecture Summary
Topic-073: Sign language: A True Language Without Speech
The production of signs is theoretically important because it allows researchers to separate cognitive processes involved in translating thought into language from the physical characteristics of the speech apparatus. Speech shares the vocal channel with respiration, whereas sign production can occur entirely in parallel with, and unimpeded by, respiratory activity. Studies of sign language production help distinguish between aspects constrained by broad biological forces and those specific to speech. A striking similarity is that errors in signing strongly resemble those found in speech production. Slips of the hand, like speech slips, typically involve a systematic error in a single linguistic unit, providing evidence that the parameters underlying signs are planned independently.
💡 Why this matters: Sign language research reveals which aspects of language are tied to the vocal-auditory channel and which reflect deeper, modality-independent cognitive organization.
Topic-074: Speech-Based Sign Languages
There are two principal types of sign language: speech-based and independent. Speech-based sign languages represent spoken words (or their spelling) and the order of words or morphemes as they appear in ordinary spoken languages like Swedish, English, and French. Speech-based sign language can be of two kinds: one representing morphemes of speech and one representing spelling (orthography). Fingerspelling translates English words letter by letter into manual form and must be learned by every signer to express proper nouns. In contrast, Independent Sign Languages (ISLs), such as American Sign Language (ASL) and British Sign Language (BSL), are not speech-based, not mutually intelligible, and have developed their own words and grammatical systems.
🔑 Definition — Independent Sign Languages (ISLs): Sign languages that are independent of ordinary spoken language, having developed their own words and grammatical systems for sentence production and understanding.
Topic-075: Basic Grammatical Concepts
Four basic grammatical concepts are discussed: duality of patterning, morphology, phrase structure, and linguistic productivity. Words are composed of phonemes, which have distinctive features, and smaller units are combined in a rule-governed manner to produce larger units. Words consist of one or more units of meaning called morphemes, and the system of grammatical morphemes allows speakers to signal subtle differences in meaning. Phrase-structure rules codify intuitions about groupings of words in a sentence, and some sentences are ambiguous because they may be grouped in more than one way. Linguistic productivity refers to the fact that there is no limit to the number of sentences in a language, with recursion being one type of phrase-structure rule responsible for this productivity.
🔑 Definition — Duality of patterning: The organization of language into two levels—meaningless units (phonemes) combine to form meaningful units (morphemes and words).
🔑 Definition — Recursion: A phrase-structure rule that allows a constituent to contain another instance of the same constituent, enabling infinite productivity.
Topic-076: Insights from Sign Language
American Sign Language (ASL) is expressed in visual or spatial form, unlike speech. ASL is sharply distinguished from manual forms of English that translate English sounds into signs. Fingerspelling translates English words letter by letter into manual form and is a secondary gestural system derived from English. ASL is independent of English and derived from French Sign Language. Although ASL was once regarded as mere pantomime or grammatically deficient, decades of research have disproven these ideas. The study of ASL enables examination of how grammatical concepts generalize to language in a visual modality.
Topic-077: Transformational Grammar
Evaluation of Grammars: Chomsky proposed three criteria for evaluating how good a grammar theory is.
Observational adequacy: The grammar must specify what is and is not an acceptable sequence in the language. This applies at multiple levels—at the phonological level, "pbort" is not acceptable; at the syntactic level, the grammar must generate grammatical sentences without generating ungrammatical strings. A grammar is observationally adequate if it generates all acceptable sequences and none of the unacceptable sequences.
Descriptive adequacy: The grammar must specify the relationships between various sequences in the language. It is not enough to mark a sequence as permissible; the grammar must also explain how it relates to other sentences that are similar or opposite in meaning.
Explanatory adequacy: This criterion addresses how linguists determine which descriptively adequate grammar is best. Chomsky's answer relates to language acquisition: children are presented with samples of language and must determine the grammar from these samples. Even though incoming data may be consistent with many grammars, children choose one particular grammar, implying that innate language constraints enable the child to deduce the correct grammar. These innate mechanisms relate to linguistic universals common to all languages.
🔑 Definition — Observational adequacy: A grammar generates all acceptable sequences in a language and none of the unacceptable sequences.
🔑 Definition — Descriptive adequacy: A grammar specifies relationships between various sequences in a language (e.g., similarity in meaning despite syntactic differences).
🔑 Definition — Explanatory adequacy: A grammar explains the role of linguistic universals in language acquisition, accounting for how children choose the correct grammar from limited input.
Topic-078: Issues in Grammatical Theory
Psychological reality of grammar: Early psycholinguistic research in the 1960s was based on transformational grammar, guided by the belief that its structures and rules were psychologically real. The derivational theory of complexity (DTC) assumed that the distance between surface structure and deep structure (measured by the number of transformations) would index psychological complexity. Early studies showed negative sentences like "The sun is not shining" were harder to comprehend than affirmative forms, but this was inconclusive since they differ in meaning. Later studies contradicted DTC—for example, "The boy was bitten" (transformationally more complex with a deletion) was not harder to comprehend than "The boy was bitten by the wolf." Similarly, particle-movement transformations showed no psychological difference. As Berwick and Weinberg (1983) noted, these results do not necessarily invalidate transformational grammar—the psychological assumptions guiding DTC may be faulty.
More recent work supports the psychological reality of linguistic theory. In passive sentences like "The dentist from the new medical center in town was invited by the actress to go to the party," the moved constituent (dentist) leaves a trace at its earlier location. Osterhout and Swinney (1993) found that comprehenders reactivate the moved noun when its trace is encountered—participants responded rapidly to words semantically related to the moved noun (e.g., "tooth" for "dentist") when presented at the trace position, but not before or after.
🔑 Definition — Derivational Theory of Complexity (DTC): The view that the psychological complexity of a sentence corresponds to the number of transformations in its derivation from deep to surface structure.
🔑 Definition — Trace: In grammatical theory, the assumed empty position left behind when a constituent moves from its original location to another position in the sentence.
The centrality of syntax: Transformational grammar has undergone changes; in Chomsky's (1995) most recent version, many specific transformational rules were replaced by broader rules like a general movement rule. However, the idea that syntactic structure is at the heart of linguistic knowledge has been controversial. Two alternative approaches are discussed.
Lexical theories of grammar place greater emphasis on individual lexical items (words). In Bresnan's lexical-functional grammar, lexical entries include meaning, spelling, pronunciation, syntactic characteristics, various forms of the word, and the different kinds of sentences each form fits. For verbs, entries include arguments or semantic roles (agent, patient) and surface structure designations (subject, object). For example, the verb kiss requires both an agent and a patient. The active form assigns agent=subject, patient=object, while the passive form (be) assigns patient=subject, agent=object of the preposition by.
🔑 Definition — Lexical-functional grammar (Bresnan, 1978): A theory emphasizing lexical items, where entries include semantic roles and surface structure assignments, reducing reliance on transformations.
🔑 Definition — Semantic roles (arguments): The roles associated with verbs, such as agent (the person doing the action) and patient (the one to whom the action is done).
⭐ Key Takeaways
Sign language is a true linguistic system that shares fundamental properties with spoken language, including systematic error patterns and hierarchical organization of units, while existing in a different modality. Independent sign languages like ASL have their own grammatical systems independent of spoken languages, challenging the view that language must be speech-based. The evaluation of grammars involves three levels of adequacy: observational (generating acceptable sequences), descriptive (explaining relationships between sentences), and explanatory (accounting for language acquisition through innate constraints). The psychological reality of grammar is supported by trace effects in sentence comprehension, though the derivational theory of complexity was largely disconfirmed. Alternative lexical theories of grammar challenge the centrality of syntax by emphasizing the role of individual words and their semantic-role assignments in linguistic structure.
🧠 Quick Revision Questions
-
What is the theoretical importance of studying sign language production compared to speech production, and what does the presence of similar error patterns in both modalities suggest?
-
Distinguish between speech-based sign languages and Independent Sign Languages (ISLs), giving an example of each type.
-
What are Chomsky's three criteria for evaluating grammars, and what does explanatory adequacy specifically address regarding language acquisition?
-
What was the derivational theory of complexity (DTC), and what evidence contradicted it? How does trace theory provide alternative evidence for psychological reality?
-
How does Bresnan's lexical-functional grammar differ from transformational grammar in its treatment of verbs and their arguments, using the example of active and passive sentences?
📘 Lecture 14 — Production of Speech and Language
📖 Overview: This lecture explores how humans produce spoken language, from initial conceptual planning to articulation. It explains the three main stages of language production—planning, grammatical encoding, and phonological encoding—and examines speech errors (slips of the tongue) as a window into these underlying psycholinguistic processes. Understanding production mechanisms is essential for grasping how the brain transforms thoughts into fluent speech.
🗂️ Topics Covered
This lecture covers the introduction to speech production stages including planning/preparation, grammatical encoding, and phonological encoding. It then examines slips of the tongue, their types (sound, morpheme, word errors), and common properties. Finally, it contrasts Freudian and psycholinguistic explanations for speech errors, detailing eight error types and their systematic patterns.
📝 Lecture Summary
Topic-079: Introduction of Production of Speech Language
Language production consists of several interdependent processes that transform a nonlinguistic message into a spoken, signed, or written linguistic signal. These processes proceed in a general order but with significant interaction between them. The first stage, planning/preparation, involves message planning—an ongoing process throughout production. Research suggests messages are planned in roughly the same order they appear in an utterance.
After identifying a message, the speaker must select individual words (lexical items) through a process called lexical selection. Words are chosen based on their meaning (semantic information), which activates the word's lemma—containing both semantic and grammatical information about the word.
The first goal of conceptual preparation is to establish which parts of conceptually available information will be encoded and in what order. The second goal is to convert conceptual information into a format suitable for linguistic formulation processes. Two key issues arise: whether conceptual preparation takes language-specific properties into account (e.g., English 'friend' lacks gender info, while Spanish requires gender marking like 'amigo' vs 'amiga'), and whether conceptual information permeates grammatical encoding beyond providing input.
🔑 Definition — Lexical selection: The process of selecting individual words (lexical items) to represent a message based on their semantic information.
🔑 Definition — Lemma: A mental representation of a word that contains both semantic and grammatical information about that word.
📌 Example — Language-specific conceptual encoding: In English, "The friend of Luis is tall" does not indicate the friend's gender. In Spanish, this must be specified: "El amigo de Luis es alto" (male friend) or "La amiga de Luis es alta" (female friend). This shows that languages differ in which conceptual properties must be realized in sentence form.
Topic-079: Grammatical Information
Grammatical encoding refers to processes involved in developing a syntactically well-formed sentence. It comprises two sub-steps: first, mapping relationships among participants (agent, patient) onto functional syntactic relations (subject, direct object) at the functional level; second, linearizing words in a language-allowed manner at the positional level.
Critical grammatical information includes the word's syntactic category (noun, verb, etc.), what objects it takes, and grammatical gender if present. Each word is assigned its grammatical and thematic role. Function morphemes (like plural /s/ or past tense /ɪd/) are added at this stage.
Distinguishing between building hierarchical and linear frames solves an important problem: speech production must be incremental for fluency, but utterances must obey language-specific word order constraints. By separating hierarchical structure construction from serial ordering, hierarchical structures can be built incrementally as lexical elements become available, then mapped to permissible linear positions. Evidence comes from studies showing that linear position of words can be primed by previous presentation of the same linear order, even when hierarchical structure differs.
🔑 Definition — Functional level processing: The first step of grammatical encoding that maps conceptual participant relationships onto functional syntactic relations (e.g., subject, direct object).
🔑 Definition — Positional level processing: The second step of grammatical encoding that linearizes words into a serial order allowed by the language being spoken.
💡 Why this matters: Separating hierarchical from linear processing explains how speakers can produce fluent, grammatically correct utterances despite planning incrementally—a key requirement for natural speech.
Topic-079: Phonological Encoding
Phonological encoding transforms the mental representation of words into a sequence of speech sounds to be pronounced. First, the phonemes of words are retrieved from the mental lexicon, together with a metrical frame specifying stress pattern and number of syllables. Following retrieval, the resulting phoneme sequence is syllabified according to language-specific syllabification rules.
The domain of syllabification is the phonological word, which can but need not coincide with a lexical word. Syllabic structure is computed on-line because actual syllabification depends on context. This explains why phonological word forms are not stored and retrieved as one entity—such a representation would need to be broken up when stored syllabification doesn't match running speech syllabification. The output forms input for articulatory processes. It remains an open issue whether transition to articulation involves accessing a syllabary—memory representations specifying motor tasks for each syllable.
🔑 Definition — Phonological encoding: The processes responsible for determining phonological word forms and prosodic content of a sentence.
📌 Example — Context-dependent syllabification: The word 'deceive' in isolation is syllabified as 'de-ceive.' In the utterance 'deceive us,' the syllable structure becomes 'de-cei-veus,' showing how running speech context changes syllabification.
Topic-080: Slips of Tongue
Slips-of-the-tongue are speech errors where intended utterances are rearranged between other words or sounds. According to psychologist Gary Dell, they are significant because they reveal a person's widespread knowledge about language, including sounds, structures, and meanings.
Three types exist: sound errors (sounds in nearby words are exchanged, e.g., "power flot" for "flower pot"), morpheme errors (morphemes are switched, e.g., "self-instruct destruction" for "self-destruct instruction"), and word errors (actual words are rearranged, e.g., "reading a dog to my book" for "reading a book to my dog"). These are also called performance errors.
According to Sigmund Freud, slips-of-the-tongue reveal unconscious thoughts and desires—called Freudian slips or parapraxes. Speech errors occur more often when speakers are nervous, tired, anxious, or intoxicated, and are common during live broadcasts due to stress. Charles F. Hockett noted that anxiety about possible lapses leads speakers to focus attention abnormally, creating "ideal breeding grounds for stuttering." Most people make up to 22 slips of the tongue per day. Errors past age 9 are called "residual speech errors" (RSEs). Speech errors may be used intentionally for humor, as with Spoonerisms.
🔑 Definition — Freudian slips (parapraxes): Speech errors that, according to Freud, reveal the thoughts and desires of the unconscious mind.
Topic-081: Types of Speech Error
Eight basic types of speech errors exist: exchanges, substitutions, anticipations, perseverations, blends, additions, deletions, and shifts. Examples are provided in the lecture table (Table 8.1).
🔑 Definition — Exchanges: Two linguistic units swap positions (e.g., "the little burst of beaden" for "beast of burden").
🔑 Definition — Substitutions: One element replaces another (e.g., substituting a similar-sounding word).
🔑 Definition — Anticipations: A later sound or word is produced earlier (e.g., "a meal mystery" for "a real mystery").
🔑 Definition — Perseverations: An earlier sound or word persists into a later position (e.g., "he pulled a pantrum" for "he pulled a tantrum").
🔑 Definition — Blends: Two competing words combine into a single non-word (e.g., blending "hiss" and "screech" into "hreech").
Topic-082: Common Properties of Speech Error
Garrett identified four generalizations about speech errors:
First, elements that interact tend to come from similar linguistic environments. Initial segments exchange with other initial segments; same for middle and final segments. Exchanges are more common when preceding segments are similar.
Second, interacting elements tend to be similar to one another. Consonants exchange only with consonants (not vowels). Errors involving similar sounds are based on phonetic similarity rather than meaning.
📌 Example — Similar environment: In "the little burst of beaden" (for "beast of burden"), the initial /b/ of "beast" and "burden" interact. In "you're not a poojin pitter-downer" (for "pigeon putterdowner"), initial /p/ sounds interact.
📌 Example — Phonetic similarity: "Sesame Street crackers" (for "sesame seed crackers") shows swapping of similar-sounding /t/ and /d/ sounds.
Topic-083: Explanations of Speech Errors
Speech errors systematically fall into eight categories: exchanges, substitutions, additions, deletions, anticipations, perseverations, blends, and shifts. Freud's view held that errors occur because we have more than one plan for production, and one plan competes with and dominates another.
More recent psycholinguistic thinking focuses on the underlying stages of production. The most common interpretation is that we produce speech through separate stages, each devoted to a single level of linguistic analysis. Errors typically occur at one level but not others during production.
Topic-084: Differences in Freudian and Psycholinguistics Explanation
The Freudian explanation: Freud emphasized psychodynamic factors making certain content more available. He argued errors "arise from the concurrent action—or rather, the mutual opposing action—of two different intentions"—one conscious, one a disturbing thought. Examples include President George Bush saying "lovely recession" (for "reception") during the 1992 campaign—possibly revealing preoccupation with the recession. A student saying "my grandmother lied" (for "died") might reveal unconscious thinking. Freud believed virtually all speech errors were caused by intrusion of repressed ideas from the unconscious.
A Psycholinguistic explanation: Modern thinking focuses on insights into language mechanisms rather than unconscious motivations. Speech errors play a role similar to aphasic disorders—providing insights into normal language functioning. Many segments that change in speech errors are precisely those postulated by linguistic theories, supporting the notion that phonetic features, phonemes, and morphemes constitute planning units during utterance production.
This view holds that we produce utterances through a series of stages, each devoted to a different level of linguistic analysis. Speech errors reveal what these specific stages look like, from idea to articulation.
📌 Example — Freudian vs. psycholinguistic interpretation: "A meal mystery" (for "a real mystery") could be Freudian (revealing hidden thoughts about food), but more simply reflects anticipation of the /m/ sound from "mystery." Similarly, "he pulled a pantrum" (for "tantrum") shows perseveration of the /p/ from "pulled."
🔑 Definition — Intrapsychic conflict: Freud's concept where two competing intentions (one conscious, one disturbing) create speech errors through mutual opposition.
⭐ Key Takeaways
Students must remember the three main stages of language production: planning/preparation (conceptual and lexical selection), grammatical encoding (functional and positional processing), and phonological encoding (phoneme retrieval and syllabification). Slips of the tongue are systematic errors that reveal linguistic structure—sound, morpheme, and word errors show that language is produced in stages with distinct processing levels. Eight error types exist (exchanges, substitutions, anticipations, perseverations, blends, additions, deletions, shifts), and Garret's four properties show they follow linguistic rules (similar environments, similar elements). The critical contrast is between Freud's unconscious-conflict explanation and the psycholinguistic view that errors reflect normal processing stages and planning units.
🧠 Quick Revision Questions
- What are the three main stages of language production, and what processing occurs in each?
- Explain the difference between functional level processing and positional level processing in grammatical encoding.
- What are the three types of slip-of-the-tongue errors, and provide an example of each?
- According to Garret's generalizations, what two properties characterize elements that interact in speech errors?
- How does the psycholinguistic explanation of speech errors differ from Freud's explanation regarding the origin of these errors?
📘 Lecture 15 — Formulating Linguistic Planning
📖 Overview: This lecture examines how speakers plan and produce spoken language, from initial idea through articulation. It covers serial models that propose step-by-step planning stages, evidence from speech errors, editing processes before and after articulation, Freud's psychodynamic interpretation of slips, parallel processing alternatives, and how subject-verb agreement errors illuminate production mechanisms.
🗂️ Topics Covered
The lecture covers serial models of linguistic planning based on Fromkin's stage theory, evidence for independence of planning units from speech error data, editing processes that monitor utterances before articulation, Freud's view of slips of the tongue as expressing unconscious conflicts, parallel models as alternatives to serial models, and the role of number agreement in evaluating production theories.
📝 Lecture Summary
Topic-085: Serial Models of Linguistic Planning
The pioneering studies of Fromkin suggested that the process of planning speech can be viewed as a series of stages, each devoted to one level of linguistic planning. This serial approach proposes that speakers move through distinct levels sequentially, beginning with the overall message, followed by syntactic organization, selection of content words, morphemes, and finally phonological encoding.
Topic-086: Independence of Planning Unit
Evidence that the stages are independent comes from speech errors: the vast majority contain mistakes at only one level of planning. For example, Fromkin's sentence (1) "singing sewer machine" (intended: Singer sewing machine) shows an error only at stage 5, where suffixes are exchanged—yet content words, stress values, and syntactic structure remain unaltered. An even more striking example is Garrett's sentence (2): "Stop beating your brick against a head wall" (intended: Stop beating your head against a brick wall).
🔑 Definition — Independence of Planning Levels: Each level of linguistic planning (semantic, syntactic, morphological, phonological) can be affected by errors while other levels remain intact, demonstrating they operate as separate processing units. 📌 Example: In "singing sewer machine," the suffix exchange (-ing and -er) does not disrupt the syntactic structure, content words, or stress patterns, showing the morphological level was the only one affected.
Topic-087: Editing Processes
In addition to planning stages, evidence indicates that editing processes intervene between the planning of an utterance and its articulation. These editing operations provide a last check to determine whether the planned utterance is linguistically and socially acceptable. While some monitoring and editing occur after speech is uttered (as we spontaneously correct ourselves), the question is whether editing also occurs prior to articulation.
💡 Why this matters: Understanding pre-articulatory editing helps explain how speakers prevent or catch errors before they are spoken, and how social and linguistic constraints shape the final output.
Topic-088: Freud’s View of Slips of Tongue
Freud emphasized the role of psychodynamic factors in speech errors. He argued that speakers have more than one idea in mind at a time, and slips "arise from the concurrent action—or perhaps rather, the mutual opposing action—of two different intentions" (Freud, 1916–1917/1963, p. 44). One intention is the conscious intention of the speaker, while the other pertains to a more disturbing thought that interferes. Sometimes the disturbing comment is censored; other times, the outcome is a slip of the tongue expressing aspects of the less conscious intention.
🔑 Definition — Freudian Slip: A speech error that reveals unconscious or repressed thoughts, occurring when two intentions (conscious and unconscious) conflict during speech production. 📌 Example 1: President George Bush said (3) "I don’t want to run the risk of ruining what is a lovely recession" (intended: reception)—possibly revealing preoccupation with the recession affecting his campaign. 📌 Example 2: A student says (4) "Last night my grandmother lied" (intended: died)—possibly revealing unexpressed thoughts about the grandmother. 📌 Example 3: A general referred to "battled scared" soldiers (intended: scarred), and a speaker said a coworker "had just expired" (intended: retired).
However, Freud's position that virtually all speech errors reflect repressed ideas is challenged by simpler errors involving anticipation (e.g., "a meal mystery" instead of "a real mystery") and perseveration (e.g., "he pulled a pantrum" instead of "he pulled a tantrum") of phonetic segments, which are better explained by cognitive processes than intrapsychic conflicts.
Topic-089: Parallel Models of Linguistic Planning
Theories of language production come in two types. Serial models assume we begin with the overall idea, followed by syntactic organization, content words, morphemes, and phonology, with errors typically involving just one level. Parallel models assume that the linguistic message is organized at semantic, syntactic, morphological, and phonological levels simultaneously. In parallel models, activation of a node at one level may trigger activation of nodes at other levels, and feedback may occur from morphological and phonological levels back to higher levels of processing. Models organized along these lines account for several important research findings.
🔑 Definition — Parallel Model: A production theory where multiple linguistic levels (semantic, syntactic, morphological, phonological) are activated simultaneously and can interact through activation spreading and feedback loops. 📌 Contrast with Serial Models: Serial models propose sequential, one-way processing where errors affect only one level; parallel models allow bidirectional activation and cross-level interaction.
Topic-090: The Role of Agreement
A line of research helpful in evaluating serial and parallel models concerns number agreement. In English, sentences require number agreement between subjects and verbs or pronouns (e.g., "The concerts this Summer have been wonderful," not "has"). Agreement errors are instructive. For example, in sentence (5) "For example, the time for fun and games is over," the head noun "time" controls the verb form, but speakers sometimes err by using "are" to match the immediately preceding word "games."
🔑 Definition — Number Agreement Error: A production error where the verb or pronoun agrees in number with an intervening noun rather than the head noun of the subject noun phrase. 📌 Example: Intended: "The time for fun and games is over" — Error pattern: "The time for fun and games are over" (matching "games" instead of "time"). 💡 Why this matters: Agreement errors help researchers test whether processing is strictly serial (following a fixed hierarchy) or parallel (allowing interference from nearby words), and reveal how syntactic structure interacts with lexical access during production.
⭐ Key Takeaways
This lecture establishes that speech production involves multiple independent planning levels, as demonstrated by speech errors affecting only one level at a time. Editing processes may occur both before and after articulation to check linguistic and social acceptability. Freud's psychodynamic explanation for slips—that unconscious conflicts intrude into speech—applies to some emotionally significant errors but fails to account for common phonological errors like anticipation and perseveration. Parallel models offer an alternative to serial models by allowing simultaneous activation and feedback across linguistic levels. Agreement errors provide a valuable empirical tool for testing these competing theories, as they reveal how nearby words can interfere with grammatical processing during production.
🧠 Quick Revision Questions
- What evidence does Fromkin's "singing sewer machine" error provide for the independence of planning levels?
- According to Freud, what causes slips of the tongue, and what examples support this view?
- How do serial models and parallel models differ in their assumptions about how linguistic levels interact during speech production?
- What is a number agreement error, and how can such errors help evaluate serial versus parallel models of production?
- What are the two types of editing processes discussed, and when in the production process does each occur?