The first 18 months of Donald Trump’s presidency have been anything but conventional.

The relentless barrage of norm-busting coming from the White House can make it difficult to keep things in perspective: Which administration actions represent major departures from the norms of U.S. democracy, which are minor deviations, and which are simply conventional actions being taken by an unconventional president?

Enter the Bright Line Watch, a group of political scientists who “monitor democratic practices, their resilience and potential threats” in the Trump era. Among their projects is a periodic survey of political science faculty at American universities, the latest wave of which just came out this week.

The survey asked 679 political scientists to rate 22 recent Trump administration actions “on their importance and on the degree to which they fit with normal political practice.” Those actions include a number of headline-grabbing events from the past few months: Trump’s face-to-face meeting with North Korea’s Kim Jong Un, his demands for greater NATO spending by U.S. allies, his repeated branding of the news media as the “enemy of the people,” among many other actions.

It’s worth noting that these results are based on subjective assessments by human beings. The scope of the inquiry is limited by the specific events the surveyors chose to ask about. But the answers provide an excellent sense of what political science scholars view as the most consequential and unconventional of those actions.

Bright Line Watch averaged the experts' responses and plotted each administration action on two axes, as seen in the chart below. The horizontal axis shows the experts' average rating of a given action’s normality (or lack thereof), while the vertical axis shows the average importance the experts assigned to each action.

The upper-left hand portion of the chart shows actions that experts viewed as both important and relatively normal: the nomination of Brett Kavanaugh to the Supreme Court, for instance, or the passage of last year’s changes to tax law. “The consequences of these policies are substantial,” as the survey report puts it, “but both are consistent with standard practice under recent administrations.”

Heading rightward into the chart, we start to arrive at actions viewed as less typical: the Dinesh D’Souza pardon and Trump’s congratulatory call to Russia’s Vladimir Putin were viewed as fairly abnormal but not particularly consequential, for instance. More substantive policy actions, like the tariffs on China and a number of American allies, were seen as similarly abnormal and more consequential.

In the experts' view, however, Trump’s most abnormal and consequential actions have been “speech acts that are unconnected to any formal presidential authority or public policy.” Last month’s joint news conference with Vladimir Putin, at which Trump appeared to endorse Russia’s denial of U.S. intelligence agencies' assessment that it had attempted to interfere in the 2016 presidential election, was ranked as highly abnormal and extremely important.

The same for Trump’s repeated criticisms of the press and FBI, and his attack on due process rights for undocumented immigrants captured at the border.

It’s reasonable to question whether statements like these, which aren’t concrete policy changes along the lines of say, separating families at the border, are truly the most abnormal or consequential acts of the Trump administration so far.

But the political scientists who run Bright Line Watch are prepared to make the case for them: “In some cases, Trump’s extreme rhetoric might expand the range of what policy options are considered reasonable or deter political opposition,” they write.

If those changes have the net effective of undermining constitutional protections or the institutions that uphold them, Bright Line Watch warns they could represent “an existential threat to American democracy.”