A student e-mailed me today & asked me to explain the differences between downward arrow and inference chaining. Also, this person wanted to know whether downward arrow could be used to get to core beliefs as well as intermediate beliefs.

It is actually kind of challenging to find a lot written about this distinction–it took me a while to see the differences myself & I got different answers from different clinicians while trying to figure it out. Some people don’t even make a distinction between the two, because they’re so similarly oriented.

Here’s what I’ve learned to date.

Downward arrow is a CT technique. The questions are asked following an inference that the client has revealed (usually in the form of an automatic thought). The questions are asked repeatedly and an arrow is drawn from one conclusion to the conclusion beneath it, over & over again (so you see a lot of downward arrows). In class, I gave 2 examples of Downward Arrow Technique questions that Aaron Beck uses, but there are many questions you could ask. Beck’s are, “What would it mean to you?” (which would generally reveal an intermediate belief — usually an assumption or rule) and the other is, “What would it mean about you?” (which could sometimes reveal a core belief if you repeat the question a couple of times).

You can also use the downward arrow technique to help with anxiety by revealing the patient’s catastrophizing. Other questions are used in this case, but the arrow is drawn below each idea, connecting it to the next one below it. So, if the patient says, “What if this will happen?” You could say, “What if it did? What is the worst thing that could happen? In the event that the worst would happen, what would you do? And then what would happen? And then what?” You’d ask until you got the patient PAST the catastrophic ideas & into the probability that he/she would cope. You could use further questions (such as what is the best that could happen? to show that it isn’t absolute that only the worst thing would happen.)

Inference chaining is a similar technique used in REBT. You could ask the more generic, “Say your inference were true. What would it mean if ___ were true?” You could get any variety of responses, ranging from a repetition of the initial inference to derivatives of musturbation (awfulizing, low frustration tolerance, or depreciation) to a demand. You would continue to ask the same generic question, using the next level of inference to fill in the blank. You then chain the inference to the next inference & continue to ask the question until you get to an inference that you think is worthy of disputation. You can dispute at any of the levels–all will usually lead to some relief.

You can also show the client how their primary inference at the top of the chain is ultimately chained to the conclusions at the bottom of the inference chain, hence showing the person how he/she creates emotional disturbance. Both of these techniques are much easier to demonstrate than to write about! I’m happy to demonstrate it in class. In both techniques, you want to lend a lot of support to the client because they are threatening for the client to endure (ego wise). I guess an analogy would be like doing surgery–if you’re removing a splinter, it hurts a little, but when you cut deeply to remove a shard of glass, it hurts a lot. So, use the strategies we discussed in class to lend support to the client as you use these technique–it is best to lend support at each point in the questioning. For example, to lend support you could say, “I know this is difficult to consider. Please understand that I’m not necessarily agreeing with your idea, however, for the sake of discussion, what would it mean if (the inference) were true?”

I am open to hearing more on this topic, so please add to the discussion.

Dr. Garcy