Facts versus Interpretations in Intelligence : A Descriptive Taxonomy for Information Evaluation

Traditionally, intelligence officers use an alphanumeric scale known as the Admiralty System to evaluate informational messages by rating the credibility of their contents and the reliability of their sources (e.g . NATO AJP-2.1, 2016). Amongst other duties, they are expected to clearly distinguish objective facts from subjective interpretations during this evaluation (NATO STANAG-2511, 2003). That being said, various experimental results show that officers are unable to properly fulfill this methodological duty (e.g . Baker et al. , 1968 ; Kelly Johnson, 1973). Our explanation is that the extant scale, which is evaluative by nature, does not allow them to endorse a more objective, that is to say descriptive, perspective on information. In this article, we aim to help enforce the facts versus interpretations recommendation in the intelligence domain. By extracting the descriptive dimensions that underlie the scale, and by grouping them by linguistic directionality (e.g . Teigen Mandel et al. , 2022), we introduce a taxonomy to categorize intelligence messages more objectively. This taxonomy is fine-grained : it integrates messages which are informative or deceptive in the classical sense (e.g . misinformation , lying) , but also more borderline messages, such as omissions and half-truths , which rely on the use of linguistic vagueness (following Égré Icard et al. , 2022). By putting descriptive lenses on information evaluation, we seek to provide new categories to help officers make more acute evaluations of information.