Abstract:

This investigation used verbal protocol methods to examine contextual vocabulary acquisition (CVA)--the deliberate, intentional, and strategic process whereby meanings of unknown words are derived from multiple natural contexts. 14 high-ability adolescent readers (5 young men and 9 young women) with reading comprehension abilities of grade-equivalent 12.9+ were invited to participate. Participants were pre-tested on knowledge of the meanings of 48 low-frequency English nouns, verbs, and adjectives. Based on pretest results, 17 target words (TW) were identified. For each TW, a set of 7-13 natural passages from newspapers, books, and magazines was assembled; each passage used the TW at least once (mean words per set = 1,278, mean passage readability = 9-10. Prior to data collection, the researcher (R) used several context clue typologies to identify text cues within each target text that suggested aspects of TW meaning. During one-to-one sessions with the R, participants were asked to read multiple authentic texts containing specific unknown TWs, to think aloud about the possible meanings of these words, and to construct dictionary-like definitions for the words.

Each participant completed between 4-8 think-aloud protocols during one-to-sessions with R. Participants read the target texts on their own and--during reading, after reading, or both--thought aloud about the possible meaning(s) of the TW. The R provided encouragement, occasionally asked probing questions, and gave meanings of non-target words when asked. At the conclusion of each think-aloud activity, the participant constructed and wrote a dictionary-like definition of the TW and rated his/her confidence in this definition. Verbatim session transcripts were analyzed qualitatively using a combination of open axial and a priori coding in order to determine the nature of the text cue moves, reasoning/inference moves, and other reader moves made by participants. The coding process was iterative and continued until no new categories emerged. 15% of the data set was independently coded. Inter-rater agreement was 80%. By talking through differences, the two raters achieved 95% agreement. Further qualitative methods were used to explore the influences of specific text, target word, and research situation factors on participants' processes and meaning-derivation outcomes.

Analysis of 74 session transcripts revealed that participants had used a similar CVA process. In general, global comprehension of texts led participants to identify specific segments of the text as cues and to reason about these segments, thereby fostering the understanding of specific target-word meanings. Prior knowledge of text topic and the vocabulary used by authors was highly facilitative of reasoning and inferencing processes and of readers' abilities to make productive use of available text cues. On average, participants appeared to notice and use fewer than half the available text cues in the target texts. Some of the target texts were unintentionally more "considerate" of readers than others in terms of the support they provided for word-meaning derivation. Yet readers were also able to derive word meanings from texts that might be characterized as "inconsiderate," because the limitations of individual texts were mitigated to some degree across multiple passages within text sets. Factors inherent to the target words (such as conceptual complexity or subtleties in intended meaning across texts) related to how much difficulty readers had constructing dictionary-like definitions, despite the facilitative influence of multiple contexts. Participants' processes were also affected by the ordering of passages in the text sets, by the verbal protocol methodology used for the research, and by the moves made by the R during the verbal protocol sessions. These findings suggest that researchers and teachers adopt broader views of "context" in CVA that includes recognition of the roles played by prior knowledge and reasoning. Additional implications for secondary reading instruction and recommendations for further research are discussed.