In the policymaking realm the framing of policy issues and problem definition have long been powerful tools in setting the agenda for decision makers and influencing the prioritization of acceptable solutions. Policy debates about online privacy are riddled with buzzwords that establish privacy as a policy domain. Some of these frames survive over time, although their meaning and interpretation change; others gain temporary visibility only to be left at the margins of the discourse; yet others disappear when new frames emerge. Over the years, and depending on venue, privacy has been framed as a matter of security, freedom of expression, human rights, and more. These discursive tokens act as boundary negotiating artifacts that delineate one policy space from another, as framing is a powerful mechanism in shaping policy debates. Frames are strategically deployed to leverage the status quo to promote one’s political agenda in policy deliberation. Taken together, the prominence and the functionality of framing make it an important factor in understanding both actual policy and its potential repercussions. Yet, frames are rarely systematically analyzed in the information policy research.This paper offers a longitudinal perspective on online privacy policy formulation through the analysis of frames used to define and refer to privacy-related concepts in a nonbinding, multistakeholder policy deliberation environment. It focuses on the consultative settings of the Internet Governance Forum (IGF) – a unique diplomatic experiment within the UN system and international policymaking; one that attracts a high level of participation from governments, industry, and civil society even though participation is voluntary and its decisions are non-binding. Some have labeled it as a decision-shaping rather than a decision-making forum or a forum where issues of Internet-related policy are being discursively constructed. Many topics discussed at the IGF are also prominent in the global public discourse, privacy being one of the most pivotal issues. This study is part of a larger project looking into how policy language is getting shaped and how it influences actual policy decisions. Working with the verbatim transcripts of IGF deliberation, we combine computational linguistics, exploratory visual analytics, and discourse analysis. Specifically, we employ a visualization of selectional preference learning, a computational technique that quantifies preferential relationships among groups of words, to identify passages where definitional framing activities may be occurring. We then apply traditional discourse analysis techniques to these passages in order to map out online privacy as a policy within the multistakeholder deliberative environment of the IGF. Our goals here are tri-fold. First, we will draw a general picture of how online privacy is being framed at the IGF. Second, we will examine how that framing has evolved over time, from the first IGF in 2006 through the sixth IGF in 2011. Third, we will look into the strategic deployment of frames by different stakeholder groups and individual influential actors in this space. We anticipate our findings to raise focused questions about privacy policy deliberation in other fora, beyond IGF, and our methodological approach to be applicable to additional Internet policy domains.