While it is well recognized that students are prone to difficulties when performing linguistic-to-algebra translations, the nature of students' difficulties remain an issue of contention. Moreover, the literature indicates that these difficulties are not easily remediated by domain-specific instruction. Some have opined that this is the case because few frameworks exist which are sufficiently robust and revelatory to provide the insights needed to address these difficulties. This study uses a model situated in both a cognitive-oriented frame and the translation process itself to analyze student activity in the linguistic-to-algebra translation process, defines error types made, and recognizes the frequencies of such.