Many types of crime are difficult to study because they are hard to operationalize, hidden from the public, or both. With communication increasingly moving to online domains, recent work has begun to examine whether the online domain contains traces of such hard-to-observe crimes. This study explores the online linguistic contours of hard-to-observe crimes through a rigorous mixed-methods approach that combines interviews and computational text analysis. Using human trafficking in illicit massage businesses as a proof-of-concept, we show how this approach, which we call semi-inductive triangulation, meets the empirical contextuality and relationality of crime traces in the online domain. The findings contribute to an emerging field of computational criminology and call for an integration of linguistic approaches in criminology.

中文翻译：

---

通过半归纳三角法识别难以观察的犯罪的在线风险标记：美国人口贩运案例

许多类型的犯罪难以研究，因为它们难以实施、对公众隐藏或两者兼而有之。随着通信越来越多地转向在线域，最近的工作已经开始检查在线域是否包含此类难以观察的犯罪痕迹。本研究通过结合访谈和计算文本分析的严格混合方法，探索难以观察的犯罪的在线语言轮廓。使用非法按摩业务中的人口贩运作为概念验证，我们展示了这种我们称之为半归纳三角测量的方法如何满足在线领域犯罪痕迹的经验背景和相关性。这些发现有助于计算犯罪学的新兴领域，并呼吁将语言学方法整合到犯罪学中。