Omar and Lena entered the university archive（档案馆）/'ɑ:kaiv/ to examine a manuscript（手稿）/'mænjuskript/ cited in Lena's dissertation（论文）/.disә'teiʃәn/. Its bibliography（参考书目）/.bibli'ɒgrәfi/ looked credible（可信的）/'kredәbl/, and the author had an impressive credential（资历）/kri'denʃәl/, but one footnote challenged the study's central premise（前提）/'premis/. The existing paradigm（范式）/'pærәdaim/ treated attention as a finite（有限的）/'fainait/ cognitive（认知的）/'kɒgnitiv/ resource. Lena asked, “Can we accept the hypothesis（假设）/hai'pɒθәsis/ when the empirical（实证的）/em'pirikәl/ evidence appears to be derived from（源自）/bi dəraɪvd frəm/ one narrow linguistic（语言的）/liŋ'^wistik/ task with a single parameter（参数）/pә'ræmitә/?” Omar answered, “No. We should draw on（利用）/drɔ ɔn/ the original records first.”

They began to compile（汇编）/kәm'pail/ every relevant excerpt（摘录）/'eksә:pt/ and build a glossary（术语表）/'glɒsәri/. A metaphor（隐喻）/'metәfә/ in the old manuscript had been translated as a technical term, although its connotation（含义）/.kɒnәu'teiʃәn/ in the surrounding discourse（论述）/'diskɒ:s/ was different. Omar rewrote the passage in concise（简洁的）/kәn'sais/ language, while Lena prepared an intelligible（易理解的）/in'telidʒәbl/ exposition（阐述）/.ekspә'ziʃәn/ of the problem in its rhetoric（修辞）/'retәrik/. Their explanation had to engage the reader's intellect（智力）/'intәlekt/, denote（表示）/di'nәut/ the literal meaning, and account for（解释）/əˈkaʊnt fər/ the historical context. A concise note can shed light on（阐明）/ʃɛd laɪt ɔn/ a complex argument when its terms are defined precisely.

Next, they inspected the numerical（数值的）/nju:'merikl/ configuration（配置）/kәn.figju'reiʃәn/. Three consecutive（连续的）/kәn'sekjutiv/ trials had been combined as if they were concurrent（同时发生的）/kәn'kʌrәnt/ and simultaneous（同时的）/.simәl'teinjәs/. That discrepancy（差异）/dis'krepәnsi/ changed the exponent（指数）/ik'spәunәnt/ in one formula and made two variables appear to correlate（存在相关性）/'kɒrәleit/. To test the effect, the team used software to simulate（模拟）/'simjuleit/ each trial separately. If one value was nominal（名义上的）/'nɒminl/ rather than measured, it could not be treated as an integral（完整的）/'intigrәl/ quantity or equate（使相等）/i'kweit/ to its reciprocal（倒数）/ri'siprәkl/. The revised calculation had to be consistent with（与其一致）/bi kənˈsɪstənt wɪθ/ the raw data.

The footnote also described an artifact（器物）/'ɑ:tifækt/ that the database had treated as an independent entity（实体）/'entiti/. Under closer scrutiny（审查）/'skru:tini/, it was merely one constituent（成分）/kәn'stitjuәnt/ of a larger instrument. Its shape could depict（描述）/di'pikt/ a separate object, but its intrinsic（本质的）/in'trinsik/ function did not signify（表示）/'signifai/ independence. This distinction mattered in terms of（就其而言）/ɪn tərmz əv/ the model's design. The team had to take into account（考虑在内）/teɪk ˈɪntu əˈkaʊnt/ every facet（方面）/'fæsit/ of the instrument and contemplate（思考）/'kɒntempleit/ whether a different configuration would give rise to（引起）/gɪv raɪz tɪ/ the same result.

Before the symposium（研讨会）/sim'pәuziәm/, Omar and Lena had to proofread（校对）/ˈpruˌfrid/ the paper, elaborate（详细说明）/i'læbәreit/ on the method, and add a complete synthesis（综合）/'sinθisis/. No conclusion could be deduced（推导）/di'dju:s/ on the basis of（基于）/ɔn ðə ˈbeɪsɪs əv/ reputation alone. Their claim would be subject to（受到约束）/bi ˈsəbʤɪkt tɪ/ the same scrutiny as any other dissertation. Limited evidence would constrain（限制）/kәn'strein/ its reach. The final article did not reject the earlier hypothesis; it showed which prerequisite（先决条件）/pri:'rekwizit/ made the model valid. That careful correction turned a small archival clue into a more credible account of cognition.