Every landlord premises case in California runs through the same choke point: the plaintiff has to show that the harm was foreseeable enough to justify the precaution the owner failed to take. That sounds like a jury question, and often it is. But duty is decided by the court, and the court decides it by asking how much foreseeability the specific demanded precaution requires. The heavier the burden you want to impose on the landlord, the more foreseeability you have to put on the table before the case reaches a jury at all.
That trade-off is not new — it traces to Ann M. and Castaneda — but the recent appellate rulings have made the calibration more explicit, and less friendly to plaintiffs who plead a general sense of a "bad neighborhood" instead of concrete prior incidents. If you handle tenant-injury and third-party-assault cases, the duty section of your opposition to summary judgment is where these cases are won or lost.
The Rowland baseline, and why foreseeability carries the weight
Start with the statute. Civil Code section 1714(a) imposes a general duty to use ordinary care to avoid injuring others, and a landlord's control over the premises brings the owner within it. Rowland v. Christian (1968) 69 Cal.2d 108 then supplies the factors a court uses to carve out exceptions to that duty: foreseeability of harm, the degree of certainty of injury, the closeness of the connection between conduct and injury, moral blame, the policy of preventing future harm, the burden on the defendant, and the consequences to the community.
In practice, most of those factors are constant across a category of cases. What moves is foreseeability weighed against burden. As the Supreme Court framed the modern duty analysis in Cabral v. Ralphs Grocery Co. (2011) 51 Cal.4th 764, courts decide duty at the level of the category of conduct, not the specific facts — so the question is whether landlords of this general type owe tenants of this general type protection against this general kind of harm. Foreseeability at that categorical level is the lever that opens or closes the door.
Ann M. to Castaneda: the heightened-foreseeability line
The sliding scale got its shape in the third-party criminal assault cases. In Ann M. v. Pacific Plaza Shopping Center (1993) 6 Cal.4th 666, the Court held that imposing a duty to hire security guards — an expensive, ongoing measure — requires a high degree of foreseeability, usually shown by prior similar incidents on or near the premises. Sharon P. v. Arman, Ltd. (1999) 21 Cal.4th 1181 applied the same rule to a parking-garage assault and found the foreseeability lacking without a record of comparable prior crimes.
Delgado v. Trax Bar & Grill (2005) 36 Cal.4th 224 then stated the principle plainly: the required degree of foreseeability slides with the burden of the precaution. A minimally burdensome step — a bar's security personnel intervening in a brewing fight they already saw — can be required on a lesser showing, even without prior similar incidents, because the cost of acting is low. A heavily burdensome step demands the prior-similar-incidents showing.
Castaneda v. Olsher (2007) 41 Cal.4th 1205 carried that framework into the landlord-tenant setting. The plaintiff argued the mobilehome-park owner should have refused to rent to, or should have evicted, a tenant with gang ties. The Court held that where the precaution demanded is that burdensome — declining or terminating tenancies based on suspected gang affiliation — the plaintiff must show heightened foreseeability, and the record did not supply it. Castaneda is the case defense counsel will cite first, and it is worth reading closely for how the Court pairs each proposed precaution with its own foreseeability threshold rather than treating "duty" as a single yes-or-no.
The low-burden counterweight: Vasquez
The sliding scale cuts both ways, and plaintiff lawyers should press the other end of it. Vasquez v. Residential Investments, Inc. (2004) 118 Cal.App.4th 269 is the model. A tenant was injured because of a missing pane of glass in a door the landlord knew about; fixing it cost roughly a dollar. The court held that where the burden of the precaution is trivial, even a modest degree of foreseeability supports a duty. You do not need prior-similar-incident proof to require a landlord to repair a known broken lock, a dark stairwell, or a defective railing.
The lesson is to frame the precaution as narrowly and cheaply as the facts allow. "Replace the burned-out bulb in a stairwell you knew was dark" is a Vasquez case. "Hire a nighttime guard patrol" is an Ann M. case. The same incident can be pled either way, and the framing decides which foreseeability threshold the court applies.
What the recent appellate work actually moved
The newer decisions have not overturned this structure — they have tightened how courts apply it, mostly at the categorical level Cabral described. Two shifts matter for landlord-tenant practice.
First, appellate panels have grown more willing to resolve the low-burden/high-burden characterization themselves rather than defer it to the jury. When a court accepts the defense framing that the plaintiff is really demanding a costly, open-ended security program, it applies the heightened standard and grants summary judgment before foreseeability ever reaches a jury. See, for example, Issakhani v. Shadow Glen Homeowners Assn. (2021) 63 Cal.App.5th 917, where the court ran the full Rowland balance and found no duty to provide additional parking or crossing safeguards for a pedestrian struck jaywalking — burden and community consequences outweighed a thin foreseeability showing.
Second, the courts have kept the door open where the landlord had actual, specific knowledge of the exact hazard. Where a landlord knew of a particular dangerous condition or a particular dangerous animal kept by a tenant, the foreseeability analysis looks very different from the diffuse "high-crime area" theory, and duty is easier to sustain. That actual-knowledge track is the reliable one; the ambient-risk track is where plaintiffs keep losing on duty.
The through-line is that generalized neighborhood danger is not doing the work it once seemed to. Courts want the foreseeability tied to something concrete — prior incidents of the same kind, a reported defect, a documented complaint — and they want it proportionate to the cost of what you say the landlord should have done.
Practice points for pleading and proof
Plead the precaution before you plead the duty. Decide what specific, low-cost step the landlord skipped, and build the duty argument around that step so you land in Vasquez territory rather than Ann M. territory. Vague allegations that the owner "failed to provide adequate security" invite the heightened standard.
Then develop the foreseeability record in discovery early. Pull prior police calls for service to the address, code-enforcement and habitability complaints, prior tenant grievances, work orders, and any correspondence showing the landlord's actual knowledge. Prior-similar-incident evidence is not only for guard-service cases; it strengthens every point on the scale. Causation is a separate hurdle you still have to clear — a self-evident defect can carry it, as the analysis in our note on the Fourth Circuit's no-expert-required slip-and-fall ruling illustrates — but do not let a strong duty theory distract from proving the mechanism of injury.
Watch the intervening-act problem in third-party assault cases. A criminal act by a stranger can break the causal chain, and defendants argue it does exactly that. The framing questions are the same ones that surface in vehicle cases where a defendant says its role was mere presence rather than cause; our breakdown of Werner v. Blake on proximate cause is a useful companion on how courts separate a condition from a legal cause. Keep the landlord's omission tied to the specific mechanism that let the harm occur — the open gate, the broken lock, the unlit path — so the third party's act reads as the foreseeable consequence the precaution existed to prevent.
Where this leaves the duty question
The California framework has not changed its architecture in a generation: duty is for the court, foreseeability is the swing factor, and the required degree of foreseeability rises with the burden of the precaution you demand. What the recent rulings have done is make the courts quicker to fix the burden characterization themselves and less tolerant of foreseeability built on neighborhood reputation alone. The plaintiff who names a cheap, specific, knowable precaution and backs it with concrete prior-incident or actual-knowledge evidence keeps the duty question open. The one who asks for a general security program on a general theory of danger hands the defense its summary judgment.