Almost exactly 20 years ago, I first wrote about federal prosecutors’ novel effort to prosecute repulsive Internet speech by using 18 USC Sec. 875(c), which prohibits “in interstate or foreign commerce any communication containing any threat to kidnap any person or any threat to injure the person of another.”

Nowadays, it’s no surprise to see entrepreneurial prosecutors still trying to apply this somewhat odd federal criminal statute to hateful Facebook postings.  But the Supreme Court decided this week that there have to be at least some limits on the scope of this seemingly broad statute.

While the court’s majority opinion (the case was decided 8-1, with Justice Samuel Alito dissenting in part and Justice Clarence Thomas dissenting entirely) stops short of invoking the First Amendment in its analysis, the decision still reflects the justices’ unwillingness to punish merely offensive speech, regardless of how infuriating or frightening it might be.

Make no mistake about it: it’s entirely understandable that the defendant Anthony Douglas Elonis’ ex-wife (the first target of his furious Facebook postings) would feel frightened at the violent implications and imagery that Elonis vented on Facebook. It’s equally understandable that the federal agents who ended up investigating Elonis would be angered at the guy’s in-your-face insistence on sounding menacing, in such a public way, even though there doesn’t seem to be evidence that he actually intended to carry out any crime.

But the court’s majority opinion pulled back from extending Section 875(c) so broadly as to require only that the alleged “threat” be language that the defendant could have foreseen would be regarded as a threat. In short, the Supremes rejected the idea that Elonis could be found guilty if he was merely negligent as to the effect of his public postings.

Instead, the high court insisted that some degree of “scienter” (usually defined as “guilty knowledge”) out to be required for conviction. (Whether the “scienter” standard is met if the defendant is reckless, rather than merely negligent, about the fear his words might cause is an issue the Supreme Court dodges, for now.)

In 25 years of doing Internet free-speech law, I’ve certainly run into trolls who set out to push freedom of speech to its limits. You probably have, too. And I respect the idea that there are some limits on what you ought to be able to say on the Internet without fear of prosecution.  The other statutes in the same chapter of the U.S. Code address some of those limits, such as issuing real threats against the U.S. president—a situation that more or less forces the Secret Service to start an investigation

But I’m still heartened by the Supreme Court’s insistence that words alone don’t create the “threat” for the purposes of 18 USC Sec. 875(c). Instead, the court writes, it’s words plus something else—in this case, evidence of some kind of “awareness of some wrongdoing” – that ought to be the baseline for prosecuting Internet speech as a crime.