I don't know of direct research on these exact beliefs, but since you've asked, I'll offer some indirect theoretical support for a positive correlation hypothesis. First, I'll reframe your belief constructs in terms of known and studied phenomena.
1) Everyone has to take care of himself; it's not mandatory for people to be nice and good, it's mandatory to defend ourselves from others and to avoid letting the others exploit us.
This statement implies advocacy of independence, an internal locus of control, security values, wariness of others' intentions, and devaluation of benevolence. Attitudes like these may relate to conservatism, authoritarianism, social dominance orientation (SDO), and low agreeableness.
2) No one is guilty if he exploits someone, since he should not let me exploit him.
This resembles the just-world hypothesis and victim blaming in general. These attitudes also relate to internal locus of control, negative evaluations of others, devaluation of universalism (which is adjacent to benevolence in value space and opposite from power values, which are adjacent to security values), conservatism, authoritarianism, SDO, and low agreeableness.
Because your two statements resemble markers of attitudes with common correlates, it is likely these attitudes correlate directly (though not necessarily without mediation by their common correlates). As for your second question, in as much as victim blaming occurs to reduce cognitive dissonance or justify behaviors to others when questioned, it is likely to occur when one is resisting social or internal pressures to empathize with the victim. It's hard (though not impossible) to be diagnosed as disordered by being too empathetic, whereas it's much easier to be diagnosed with disorders like psychopathy and antisocial personality disorder for lack of empathy. Hence if one blames a victim to justify non-empathetic behavior to oneself or others, one is more likely to behave this way in the future, and more likely to be seen as lacking empathy by others. Yes, it stands to reason that one would increase one's chances of being diagnosed with a disorder by displaying such a symptom of empathy-deficient disorders...but this doesn't mean that believing the just-world hypothesis increases the likelihood of becoming a psychopath.
Being a psychopath probably increases the likelihood of believing the just-world hypothesis, but clinical definitions of psychopathy do not generally conceptualize psychopathy as caused by beliefs. Diagnoses may be caused (partially) by beliefs, but where the true disorder exists, it probably has other causes. However, psychopathy as a continuous personality trait might be affected by victim blaming. Since this definition is somewhat less strict, and it's probably even normal to score above the minimum on given measures of psychopathic personality dimensions, it's somewhat less remarkable if (i.e., more likely that) a belief affects these scores. For instance, it's not hard to imagine an experiment achieving an effect on psychopathic trait dimension scores by priming the just-world hypothesis. Self-perceptive processes and spreading activation of semantic networks that link various beliefs and attitudes marked by low empathy could easily explain such effects. For example, consider this experimental priming research design:
- Participants are randomly sampled from the same population and randomly assigned to experimental and control groups. Both read vignettes about a homeless person being violently mugged, inducing strong emotions and cognitive dissonance.
- The experimental group's vignette presents the victim's background in a culpable light, such as by describing the victim's lack of effort to graduate from high school, get a job, control his alcohol addiction, stay away from dangerous people, and accept help from relatives and public support systems.
- The control group's vignette presents the victim's background in a neutral, relatively character-irrelevant manner, as by describing his favorite foods, colors, place of birth, normal performance in high school, and maybe a few normal relationships with others.
- Both groups complete measures of psychoticism and rate agreement with various statements of the just-world hypothesis and general victim blame (as a manipulation check).
- Group differences in psychoticism and its correlation with levels of just-world hypothesis endorsement are calculated. Hypothetically higher psychoticism in the experimental group and a positive correlation with just-world hypothesis endorsement, if found, are interpreted as support for the theory that just-world hypothesis belief and victim blaming attitudes induce psychotic states.
Participants could complete the same measures as in step 2 at a later time to assess lasting effects on trait-level psychoticism. Other measures with established convergent validity could also be introduced at later assessments (and probably administered first) to help rule out effects of familiarity with the original measures and memory of one's previous responses. Stronger manipulations would likely be required to produce reliable group differences at later assessments, but if not, this would be a pretty important result. Someone should probably try this experiment!
Of course, it's designed to be falsifiable, so a lack of support or even counterevidence could also result, and positive results would need to be interpreted cautiously. This design would only produce direct evidence of any effects on psychoticism, not psychopathy per se. There's also reason for ethical concern here, as one certainly wouldn't want to experimentally induce lasting changes in psychopathy, which is related after all. However, with a strong debriefing process that:
- explains the moral relevance of the manipulation
- preaches caution about endorsement of the just-world hypothesis
- emphasizes that the experimental vignette was completely fictional and intended to reflect real stereotypes about homeless people that are often unfounded and socially harmful, and
- refers participants to the right mental health support resources if they are distressed, and maybe some charity or volunteer organizations that support the local homeless population if they are interested...
I think this design could pass institutional review.