As another poster noted, too, it also helps to have a 'buddy'. If you can find a buddy (or if you have one, a spouse, partner, or close friend) who will hold you to it, call or email your buddy and say "Hey buddy, in 48 hours I will be sending my article on X to the journal of Y. Please email me or call me in 48 hours to check that I did it." Your buddy has to be willing to be stern with you so that you take the accountability seriously.
A diabolical technique I've heard of was to choose a charitable organization that fights for the opposite side of an issue you feel passionate about [eg. if you were a neo-Nazi you might choose the Southern Poverty Law Center ;-) ]. Give $50 to a friend and tell your friend that if you have not revised and submitted your article to journal Y by X date to donate that $50 to the organization that you are against. I heard that it worked in an experiment on trying to get scholarly writers to produce more pages.
There is a website where you can write in and set your goals, and then if you don't meet them, the website automatically sends a pre-set amount of money to some organization. The most effective is indeed to send money to a hated organization. Of course, you have to be honest about your goals and whether or not you've met them. I can't think of the name of the website, but someone on the fora must.
Found the website where you commit to your goals by "signing" a virtual "contract," and if you fail (you indicate a referee), they donate your money, especially effective if it's to a hated organization. See
http://www.stickk.comOther, less drastic options include
http://www.pledgehammer.com (honor system, you just make a pledge to fulfill your goals or else donate to some charity)
http://www.mygoals.com (email reminders of your goals)Hope this helps!
I feel for you OP - I'm in a similar position. I was wondering whether anyone has used mygoals.com, and if they'd recommend it? It sounds a great idea, but I wonder whether it would actually just become another distraction, and be easily ignored.