Physicists may need to tweak what they think they know about Werner Heisenberg’s famous uncertainty principle.

Measuring light particles doesn’t push them as far into the realm of quantum fuzziness as once thought, new research suggests. The work doesn’t invalidate the principle underlying all of modern quantum theory, but may have implications for supersecure cryptography and other quantum applications.

“The real Heisenberg uncertainty principle is alive and well,” says Lee Rozema, a graduate student at the University of Toronto whose team reports the finding in the Sept. 7 *Physical Review Letters*. “It’s really just this [one aspect] that needs to be updated.”

In its most famous articulation, Heisenberg’s uncertainty principle states that it’s possible at a given moment to know either the position or momentum of a particle, but not both. This relationship can be written out mathematically. But Heisenberg first came up with the idea in a slightly different fashion using slightly different mathematics. That version says the more you disturb a particle, the less precisely you can measure a particular property of it, and vice versa.

As an example, Heisenberg imagined shining particles of light on an electron and, by watching how the light bounced off it, deducing the position of the electron. But each time the light particles impart a little of their momentum to the electron, thus blurring how well scientists can measure the system. “This is how Heisenberg thought, but it wasn’t what was rigorously proven later,” says Rozema. “Physicists quite often confuse the two.”

Heisenberg’s original version still works for the light/electron example, Rozema says, but not in more general cases — as most scientists have assumed.

In 2003, Japanese physicist Masanao Ozawa showed mathematically that Heisenberg’s first version couldn’t be right. Earlier this year, he and a research team at the University of Vienna reported lab experiments confirming this.

Now, the Toronto physicists have weighed in with what they call a more direct measurement. They took single light particles, or photons, and measured two directions in which the light waves oscillated. The first measurement was a “weak” probe, gently inquiring about oscillations in one direction and then the other. Then the scientists made a “strong” measurement, directly probing whether that first, weak measurement had disturbed the system.

By combining the weak and strong measurements, Rozema’s team showed that the measured oscillations did not fit the mathematics of Heisenberg’s first formulation of the uncertainty idea. In other words, shrinking the inaccuracy of a particle measurement (making it more precise) doesn’t disturb the particle quite as much as scientists had thought.

“It is possible for both the inaccuracy and the disturbance to be small, although not both strictly zero,” says Howard Wiseman, a physicist at Griffith University in Brisbane, Australia, who proposed the measurement the Toronto team used.

The discovery is important for anyone trying to build an unbreakable quantum code. Quantum cryptography relies on the fact that eavesdroppers would be spotted by the disturbance they make. If the disturbance is smaller than expected, then eavesdroppers might be harder to detect.

“The new relation will open up new science and technology in the field of quantum information,” says Ozawa, now of Nagoya University. “It also presents a profound philosophical problem.”