Intro to Semantics and Pragmatics Unit 15 ReviewExperimental Approaches to Meaning

Key Concepts and Terminology

• Semantics studies the meaning of words, phrases, and sentences in a language
• Pragmatics focuses on how context and speaker intentions influence meaning
• Experimental semantics and pragmatics use empirical methods to test theories about meaning
  • Involves collecting data from human participants in controlled settings
  • Aims to provide objective evidence for or against specific hypotheses
• Key terms include variables (independent and dependent), stimuli, conditions, and statistical significance
• Compositionality principle states that the meaning of a complex expression is determined by the meanings of its parts and how they are combined
• Gricean maxims describe how cooperative conversation works (quantity, quality, relation, manner)

Theoretical Background

• Formal semantics uses logic and model theory to represent meaning
  • Focuses on truth conditions and entailment relations between sentences
• Cognitive semantics views meaning as conceptual structures in the mind
  • Emphasizes embodied experience, metaphor, and mental spaces
• Relevance theory in pragmatics argues that utterances raise expectations of relevance
  • Hearers interpret meaning based on maximizing relevance with minimal processing effort
• Speech act theory classifies utterances by their intended communicative function (assertives, directives, commissives, expressives, declarations)
• Experimental approaches aim to empirically test predictions made by these theories
  • Seek converging evidence from multiple methods (reaction times, eye-tracking, neuroimaging)

Experimental Design Basics

• Formulate a clear research question and hypothesis based on theoretical predictions
• Identify relevant variables to manipulate (independent) and measure (dependent)
  • Independent variables are the factors being tested (word meaning, context, speaker intention)
  • Dependent variables are the outcomes measured (reaction time, accuracy, brain activity)
• Create stimuli that isolate the variable of interest while controlling for confounds
  • Use minimal pairs differing in one key aspect (literal vs. metaphorical, sincere vs. sarcastic)
• Determine appropriate task for participants to perform (judgment, production, comprehension)
• Choose between within-subjects (each participant does all conditions) and between-subjects (different groups per condition) designs
• Decide on sample size and characteristics (age, language background) for sufficient statistical power

Data Collection Methods

• Behavioral tasks measure overt responses from participants
  • Lexical decision task asks if a string of letters is a real word or not
  • Sentence verification task asks if a sentence is true or false given a scenario
  • Self-paced reading records time spent on each word in a sentence
• Eye-tracking records gaze patterns while reading or viewing scenes
  • Fixation durations and saccades can indicate processing difficulty or anticipation
• Neuroimaging techniques measure brain activity during language processing
  • Event-related potentials (ERPs) show millisecond-level changes in electrical activity
  • Functional magnetic resonance imaging (fMRI) localizes blood flow to specific brain regions
• Collecting data online allows for larger and more diverse samples
  • Platforms like Amazon Mechanical Turk and Prolific facilitate participant recruitment and payment

Analysis Techniques

• Preprocessing data involves filtering out incorrect responses and outliers
  • Reaction times more than 2-3 standard deviations from the mean are often excluded
• Descriptive statistics summarize key properties of the data
  • Mean, median, standard deviation, standard error
• Inferential statistics test for significant differences between conditions
  • T-tests compare means between two groups
  • ANOVAs test for effects of multiple factors and their interactions
  • Regression models can include continuous predictors and covariates
• Effect sizes indicate the magnitude of differences (Cohen's d, eta-squared)
• Data visualization through graphs and plots helps identify patterns
  • Bar plots for categorical variables, scatterplots for continuous variables

Interpreting Results

• Determine if results support or fail to support the initial hypothesis
  • P-values less than .05 indicate statistically significant differences
  • Confidence intervals estimate the range of plausible values for an effect
• Consider alternative explanations and potential confounds
  • Could differences be due to low-level factors (word frequency, length) rather than meaning?
• Relate findings back to theoretical debates and previous literature
  • Do results converge with or diverge from prior studies?
  • What new questions or predictions do they generate?
• Acknowledge limitations in generalizability and ecological validity
  • Experiments may not fully capture real-world language use
  • Samples may be skewed toward certain demographics (college students, English speakers)

Challenges and Limitations

• Balancing experimental control with naturalness of stimuli
  • Highly controlled stimuli may lack context and pragmatic richness
  • More naturalistic stimuli may introduce confounds and variability
• Avoiding task demands and strategies that obscure underlying processes
  • Participants may rely on surface-level heuristics rather than fully processing meaning
• Accounting for individual differences in language processing
  • Factors like working memory, language proficiency, and cultural background can modulate effects
• Integrating findings across different methods and modalities
  • Behavioral, eye-tracking, and neuroimaging data may not always converge
  • Different tasks may tap into different aspects of meaning processing
• Dealing with null results and replication failures
  • Non-significant findings can be difficult to interpret and publish
  • Direct replications are necessary to establish robust effects

Real-World Applications

• Improving natural language processing systems that can handle ambiguity and context
  • Chatbots, virtual assistants, and machine translation rely on semantic and pragmatic knowledge
• Diagnosing and treating language disorders that affect comprehension
  • Aphasia, autism, and schizophrenia can involve deficits in processing meaning and intent
• Enhancing education and assessment of linguistic skills
  • Experimental findings can inform language teaching methods and materials
• Facilitating cross-cultural communication and understanding
  • Recognizing differences in pragmatic norms and expectations across languages and cultures
• Developing more persuasive and effective messaging in advertising, politics, and health campaigns
  • Experimentally testing how wording and framing influences interpretation and behavior