The data set ------------ This data set contains 1208 speech errors from English conversations, including 1094 lexical substitution errors and 114 word blends. It is described in detail in the first publication below and drawn from the larger SFUSED English database. If used in a new publication, please cite the data set appropriately. Alderete, John, Baese-Berk, Melissa, Brasoveanu, Adrian, and Jess H. K. Law. 2023. A new corpus of lexical substitution and word blend errors: Probing the semantic structure of lemma access failures. Manuscript, Simon Fraser University, Oregon University, and UC Santa Cruz. Alderete, John. (2019). Simon Fraser University Speech Error Database - English Beta, [www.sfu.ca/people/alderete/sfused]. Department of Linguistics, Simon Fraser University (Distributor). Files ----- **1208 lexsubs-blends_21 variables.xlsx**: Excel spreadsheet with 1208 rows (individual speech errors) and 21 columns (variables, or fields). **1208 lexsubs-blends_21 variables.csv**: comma delimited spreadsheet with 1208 rows (individual speech errors) and 21 columns (variables, or fields). **1208 lexsubs-blends_21 variables.twb**: a Tableau workbook for with all the data and worksheets for each of the analyses in the article; can be explored by using the 21 variables as dimensions and count (i.e., number of speech errors) as a measure. Field descriptions ------------------ **recordErrorID**: unique number for the speech error, enabling lookup in the database **exampleIntended**: the spelled intended word **exampleIntendedIPA**: the intended word in IPA transcription **exampleErrorSpelled**: the spelled error word **exampleErrorIPA**: the error word in IPA **exampleExample**: the longform of the speech error, complete with its linguistic context **exampleCorrected**: was the error corrected by the speaker, y = yes **exampleNotes**: any notes made by the data collector or analyst of interest, including any special assumptions needed to analyze the error **classMasterType**: the type of error, distinguishing lexical substitutions and word blends **classLevel**: Phrase for error on a syntactic phrase, Word for word-level error **classContextual**: Y if error is contextual (i.e., the error word comes from context), N if noncontextual **classMasterTypeAlternate**: alternate mastertype of the stated mastertype; alternative both inferior analysis of the speech error **wordPosIntended**: part of speech of the intended word **wordPosError**: part of speech of the error word **wordSemanticRelationship**: one of six semantic relationships between the intended and error word (see article) **wordSemanticFrame**: the semantic frame for the stated semantic relationship (see article) **wordAgreeTopSemRelYN**: (for lexical substitutions only) was there agreement between the two classifiers of the best (top) analysis of the semantic relationship (see article), Y = yes, N = no **wordAgreeAnySemRelYN**: (for lexical substitutions only) was there agreement between the two classifiers of any analysis of the semantic relationship (see article), Y = yes, N = no **specialSameInitialSegment**: do the intended and error words start with the same phonological segment? Y = yes, N = no **specialSemanticallyRelated**: are the intended and error words semantically related (see article): Y = yes, N = no **specialNontrivialFormResemblanceYN**: is there a non-trivial resemblance between the form of the intended word and the form of the error word (see article for method): Y = yes, N = no