MSc projects and latterly our MSc program, both of which needed all of her customary tact, patience, and persuasion. For five years Alison kindly maintained the illusion that I managed her. Contrariwise, I depended on her strongly, as a shoulder to cry on, someone to moan at and to gossip with, someone with whom to share hard decisions. I'll remember a kind and gentle person, a fine colleague and an engaging friend.

We offer our warmest sympathy and solace to Alison's family: Her beloved husband Richard, children Sophie and Andrew, parents Kay and David, brothers Chris and Richard, to Richard's family, and to her close friends.

In a very short time Janet identified tense and aspect, and temporal expressions in general, as her area of interest—an area that would remain the main focus of her research throughout her career. She went through her Ph.D. career very quickly, the first among her Linguistics colleagues to pass quals, the first to publish a paper at a major linguistics conference (Hitzeman 1991), and the first to graduate, in 1993, with a dissertation on "Temporal Adverbials and the Syntax-Semantics Interface" in which she proposed that the alternative interpretations of temporal adverbials like for could actually be derived from information structure without the need to stipulate an ambiguity (Hitzeman 1997).

When she finished her Ph.D., Marc Moens, then head of the Language Technology Group in Edinburgh, offered her her first job. Here is Marc remembering how this came about:

I... interviewed Janet for the LTG job in a car in Rochester, driving to Len Schubert's house for an ACL Program Committee party. It was a cold January evening. I realized pretty soon that it would be great to have her join the LTG. At the end of the interview, she parked the car next to what turned out to be a ditch filled with snow. As I got out of the car, I sank to my knees in the snow and couldn't move. Janet looked at me and said, "And I thought I was doing so well with the other interview questions."

The snow didn't change Marc's mind, and Janet spent six years in Edinburgh, continu­ing her work on temporal expressions but also starting research in new areas including discourse, natural language generation, and prosody. Her work on temporal expression interpretation led to her most cited article (Hitzeman, Moens, and Grover 1995), also included in The Language of Time, the recent collection of classics in temporal expression interpretation edited by Mani et al. (Mani, Pustejovsky, and Gaizauskas 2005). Her work on the ILEX NLG project (Hitzeman, Mellish, and Oberlander 1997) resulted instead in her getting interested in discourse, in particular the effect of local and global salience on the interpretation of anaphoric expressions (Hitzeman and Poesio 1998; Poesio et al. 2004). This research was grounded in empirical data resulting from an annotation effort that Janet started and that was to lead eventually to the creation of the GNOME corpus. She also became interested in the relation between discourse structure and prosody, getting involved in one of the first projects attempting to use discourse information to improve speech synthesis (Hitzeman et al. 1998). Finally, she became a key element of the social life at the Human Communication Research Centre, which had at its heart the Friday pub night at Maxie's that she instituted and drove forward with Matt Crocker and Shona Douglas. Here is Matt reminiscing about that time:

I remember those Maxie's evenings very fondly ... indeed, one of the most positive associations I have with my time in Edinburgh. Very often those evenings began with Janet, Shona [Douglas], and I (the hard-core Maxie's team) sharing a bottle of Montepulciano, as the rest of the crowd slowly trickled in . . . Great times.

In 2000 Janet moved back to the U.S. to work at MITRE Bedford, where she continued her research on corpus annotation, discourse, and temporal expressions but moved the primary application focus to information extraction.

At MITRE, Janet was a resident expert on the design and development of annotation schemes, and was happiest when up close to the linguistic data, especially non-garden-variety corpora such as financial filings, disease-tracking e-mails, chat room logs, blogs, and so forth. In her early years at MITRE Janet contributed to work on developing event extraction systems, working largely with rule-based mechanisms that exploited her understanding and interests in how language could express information of relevance to various tasks. A hallmark of her work was the close interaction of annotated corpus development and the design of heuristics grounded in that data. She contributed to the early versions of standards like TimeML, and was a key figure behind the development of SpatialML.

Janet was a key contributor in the development of cutting-edge natural language processing applications in such diverse areas as global disease outbreak monitoring, topic tracking in chat, and fraud detection from financial data. Among the best known papers resulting from Janet's work on information extraction while at MITRE are Mani et al. (2008) on SpatialML and Wellner et al. (2007) on de-identification. At the time of her death, Janet was in her second year as principal investigator of a project developing a prototype to monitor World Health Organization disease outbreak announcements. This system was able to exploit the unique aspects of this "micro-genre" to achieve very high levels of accuracy in automatically recognizing and mapping disease outbreak information with Google Earth (Hitzeman 2008). Janet also got involved in research on using dialogue for collaborative learning, in cooperation with Brad Goodman and colleagues (Goodman et al. 2005). Last but not least, she developed large portions of hand-annotated corpora still used to pursue research of various kinds: In addition to the SpatialML data, she was the key annotator for the cross-document coreference annota­tion corpus developed for the 2007 Johns Hopkins Summer Workshop on coreference resolution (Day et al. 2008).

Janet gained the respect of a wide variety of staff in the company. She was sought after not only for her expertise and diligence but as one of the most considerate and pleasant of colleagues to work with, always at hand with a wry comment or an astute linguistic observation. She managed to continue contributing to the area of the linguis­tics of temporal expressions, even as her MITRE work diverged into many different areas.

Janet was always at the center of social activities at conferences like ACL and LREC, and at summer schools like ESSLLI, and was popular with everybody in the community. This was because she was incredibly warm and always had a charming interest in what people were up to, in research and in personal life. An additional reason for her popularity was that she was a great storyteller and a great listener. Her e-mails were always eagerly received as they would always contain consistently funny accounts of her encounters with Superglue, dogs, or camels. In a word, she was a great source of joy for all of us.