Previous research into conditional inducements has shown that readers are sensitive after reading such conditionals to pragmatic scope differences between promises and threats; specifically, threats can be referred to as promises, but promises cannot be referred to as threats. Crucially, previous work has not revealed whether such scope effects emerge while processing the conditional itself. In the experiment reported here participants’ eye-movements were recorded whilst they read vignettes containing conditional promises and threats. We observed a reading time penalty on the conditional itself when participants read a conditional promise that was described as a “threat” (e.g., Liam threatened Perry "if you tell dad, then I'll take equal responsibility"). There was no such penalty when the word "promise" was presented before a conditional threat. These results suggest that readers are sensitive during reading of the conditional itself to pragmatic scope differences between “threats” and "promises".