If you use the word "weekend" in a workplace email, chances are you are sending the message up the organizational chart. The same is true for the words "voicemail," "driving" and "okay." Certain words and phrases are reliable indicators of whether workplace emails are sent to someone higher or lower in the corporate hierarchy, according to a new study by Georgia Tech.

Eric Gilbert, PhD, assistant professor in the School of Interactive Computing, focused his attention on the "Enron corpus," a body of 500,000 emails among about 150 former Enron employees. It is the largest email dataset available for public study.

Even after taking a conservative approach to the process by applying numerous filters and eliminating thousands of emails that would have muddied the conclusions, he still was able to identify lists of words that reliably predicted whether emails traveled up or down the ladder.

"Across a wide variety of messages and relationships, these phrases consistently stand out as signaling a power relationship between two people," Dr. Gilbert said. "The probability of it occurring due to chance alone is less than one in 1,000."

His work could be applied in designing smarter email software. Future email clients might be able to differentiate between emails sent from superiors or subordinates and then use that information to better address a person's email preferences. For example, messages sent after 5 p.m. from people under you might get held for delivery until the next day, while emails from a superior or the person's assistant would go through right away.

"We have organizational charts, but they don't tell the whole story," said Dr. Gilbert, adding that the research could help map "informal power and reporting structures" in an organization. "A classic example is the CEO's administrative assistant. That person may not occupy a high box on the org chart, but he or she still has a large amount of influence."

The top five words or phrases that are "upward predictors" are "the ability to," "I took," "are available," "kitchen," and "thought you would." The top five "downward predictors" are "have you been," "you gave," "we are in," "title," and "need in."

Dr. Gilbert compiled 100-word lists that reliably predict hierarchical direction. They can be accessed in his paper at http://comp.social.gatech.edu/papers/cscw12.hierarchy.gilbert.pdf.

The project falls in the research field of applied natural language processing, which has been an important subfield of artificial intelligence for the past 25 years. Tools developed in this area can be applied to other data and yield enlightening results about human interaction through electronic media or simply how people relate to each other.