Language Rules

Communication guidelines for therapy data systems
How to speak about mental health safely and supportively

Overview

Language matters enormously in mental health contexts.

The words you choose can:

• ✅ Validate and support users
• ❌ Pathologize normal emotions
• ❌ Trigger shame or distress
• ❌ Create false authority

This guide provides specific language rules for therapy data systems.

Core Principle

Speak as a supportive companion, never a clinical authority.

You are observing patterns alongside the user, not diagnosing from above.

Rule 1: Observation Over Diagnosis

❌ Diagnostic language (forbidden):

• "You have depression"
• "You're anxious"
• "This is a panic attack"
• "You suffer from bipolar disorder"
• "Your symptoms indicate PTSD"

✅ Observational language (required):

• "You've been feeling down lately"
• "You mentioned feeling anxious"
• "That sounds really overwhelming"
• "You've noticed big mood swings"
• "You've described some difficult experiences"

The difference:

• Diagnostic language claims medical authority
• Observational language reflects what the user has shared

Apply everywhere:

• UI copy
• Notifications
• Prompts
• Insights
• Reports

Rule 2: Person-First Language

❌ Identity-based (avoid):

• "You're depressed"
• "Anxious users"
• "Bipolar individuals"
• "A depressive"

✅ Person-first (use):

• "You've been feeling depressed"
• "People experiencing anxiety"
• "People with bipolar disorder" (when referring to diagnosed individuals)
• "Someone experiencing depression"

Why: People are not their conditions. Language should reflect full personhood.

Exception: Some communities prefer identity-first (e.g., "autistic person" vs. "person with autism"). When referring to specific communities, respect their preference. For general communication, default to person-first.

Rule 3: Normalize Emotions, Not Disorders

❌ Pathologizing normal emotions:

• "Your sadness is abnormal"
• "Feeling anxious isn't normal"
• "This level of emotion is concerning"

✅ Normalizing human emotion:

• "It's okay to feel sad"
• "Anxiety is a common human emotion"
• "Everyone experiences difficult feelings"

But also:

• Know when to suggest professional help
• Don't minimize persistent, severe symptoms
• Balance normalization with appropriate referrals

Example:

if (lowMoodDays < 3) {
  return "It's normal to have down days. Be gentle with yourself.";
}
 
if (lowMoodDays >= 14) {
  return `
    You've been feeling low for a while. While difficult emotions are normal,
    persistent low mood might benefit from professional support.
    Consider talking to a mental health professional.
  `;
}

Rule 4: Validate Feelings

❌ Dismissive language:

• "Just think positive"
• "Others have it worse"
• "You're overreacting"
• "Snap out of it"
• "It's not that bad"

✅ Validating language:

• "That sounds really difficult"
• "Your feelings are valid"
• "It makes sense you'd feel that way"
• "I hear that this has been tough for you"

Validation doesn't mean agreement. It means acknowledging the person's emotional experience.

Examples:

Rule 5: Suggest, Don't Prescribe

❌ Prescriptive (authoritative):

• "You need therapy"
• "You should try CBT"
• "You must talk to a doctor"
• "Stop doing [behavior]"

✅ Suggestive (supportive):

• "Talking to a therapist might help"
• "Some people find therapy helpful"
• "Consider consulting a mental health professional"
• "You might find it helpful to..."

The difference:

• Prescriptive language assumes authority
• Suggestive language offers options

User autonomy is critical. They make decisions about their care.

Rule 6: Avoid Clinical Jargon

❌ Clinical terms (avoid):

• "Major depressive episode"
• "Generalized anxiety disorder"
• "Comorbidity"
• "Psychopathology"
• "Differential diagnosis"

✅ Plain language (use):

• "Persistent low mood"
• "Ongoing anxiety"
• "Multiple challenges"
• "Mental health"
• "Understanding what's happening"

Exception: When quoting professional resources or educational content, clinical terms are okay if explained in plain language.

Example:

"If you're diagnosed with major depressive disorder (persistent depression that affects daily life), your doctor may recommend..."

Rule 7: Crisis Language = Immediate Resources

❌ Therapeutic response to crisis:

• "Tell me more about these feelings"
• "I'm here for you"
• "Let's work through this together"

✅ Immediate external referral:

🚨 If you're in crisis or considering self-harm:

• Call 988 (Suicide & Crisis Lifeline)
• Text HOME to 741741 (Crisis Text Line)
• Call 911 for emergencies

These resources are available 24/7.

No other response during crisis detection.

Crisis keywords:

• Suicide, kill myself, end my life
• Self-harm, hurt myself, cut myself
• Not worth living, want to die
• Overdose, pills, hanging, gun (in context)

Always: Display resources prominently, clickable/callable, no shame language.

Rule 8: Respect Privacy and Anonymity

❌ Sharing assumptions:

• "Tell your therapist about this app"
• "Share your insights with your doctor"
• (Implying they're in treatment when they might not be)

✅ Respecting privacy:

• "If you're working with a therapist, you might want to share..."
• "Some people find it helpful to discuss patterns with their doctor"
• (Offering options without assumptions)

Never assume:

• User is in therapy
• User has told anyone about their mental health
• User wants others to know

Always:

• Respect that mental health is private
• Give user control over data sharing
• Avoid pressure to disclose

Rule 9: Inclusive, Non-Discriminatory Language

❌ Exclusionary language:

• Assuming gender, relationship status, ability, etc.
• "Your husband/wife" (assume heterosexual relationship)
• "Walk it off" (assumes physical ability)
• Cultural insensitivity

✅ Inclusive language:

• "Your partner" (gender-neutral)
• "Take a break" or "Go easy on yourself" (accessible)
• Culturally aware (understand different expressions of distress)

Mental health is universal. Language should be accessible to all.

Rule 10: Strengths-Based Framing

❌ Deficit-focused:

• "You failed to journal today"
• "Your mood is bad again"
• "You're struggling"

✅ Strengths-focused:

• "You've journaled 5 days this week!"
• "You've been tracking your mood consistently"
• "You're working on understanding yourself"

Acknowledge effort and growth, even in difficult times.

Example:

if (journalStreak >= 5) {
  return "You've been so consistent with journaling. That takes real commitment.";
}
 
if (journalStreak === 0) {
  return "Ready to journal when you are. No pressure.";
  // NOT: "You haven't journaled in a while. Try to be more consistent."
}

Practical Examples

Feature: Mood Insights

❌ Bad:

Your mood has been low for 2 weeks. You're showing signs of depression.
You need to see a therapist immediately.

✅ Good:

You've been feeling low for a couple of weeks. Persistent low mood can be tough.

Consider talking to a mental health professional who can provide support.
In the meantime, be gentle with yourself.

Feature: Journal Prompt

❌ Bad:

Tell me about your childhood trauma and how it's affecting you today.

✅ Good:

What's one positive moment from today, no matter how small?

Feature: Notification

❌ Bad:

You haven't logged your mood today. Your depression tracking won't be accurate.

✅ Good:

Check in with yourself today 💙
(No pressure—just a gentle reminder)

Feature: Crisis Detection

❌ Bad:

[User mentions suicide]

I'm detecting that you're at high risk. Let me help you work through this.

✅ Good:

🚨 If you're in crisis or considering self-harm:

• Call 988 (Suicide & Crisis Lifeline)
• Text HOME to 741741 (Crisis Text Line)
• Call 911 for emergencies

These resources are available 24/7.

Feature: Professional Referral

❌ Bad:

You have severe depression. You must see a psychiatrist for medication.

✅ Good:

You've been experiencing persistent low mood for a while.

Consider reaching out to a mental health professional:
• Find a therapist: psychologytoday.com/us/therapists
• Free/low-cost options: openpathcollective.org

They can provide personalized support.

Testing Language

Automated checks:

const FORBIDDEN_PHRASES = [
  "you have depression",
  "you're bipolar",
  "you suffer from",
  "this is a panic attack",
  "you need therapy",
  // ...hundreds more
];
 
function validateCopy(text: string): boolean {
  return !FORBIDDEN_PHRASES.some(phrase =>
    text.toLowerCase().includes(phrase)
  );
}

Human review:

• Mental health professionals review all copy
• Test with diverse users
• Solicit feedback on tone and language
• Iterate based on sensitivity feedback

Language Validation Checklist

Before shipping UI copy, verify:

• Observational, not diagnostic
• Person-first language (when appropriate)
• Normalizes emotions without minimizing
• Validates feelings without being dismissive
• Suggests, doesn't prescribe
• Plain language, avoids jargon
• Crisis language triggers immediate resources
• Respects privacy and autonomy
• Inclusive and non-discriminatory
• Strengths-based framing

Summary

Language rules for therapy data systems:

1. Observation over diagnosis — "You've been feeling down" not "You have depression"
2. Person-first — "People experiencing anxiety" not "anxious people"
3. Normalize emotions — But refer serious symptoms to professionals
4. Validate feelings — "That sounds hard" not "Just think positive"
5. Suggest, don't prescribe — "Might help" not "You need"
6. Avoid jargon — Plain language over clinical terms
7. Crisis = resources — 988, Crisis Text Line, 911 (no therapeutic response)
8. Respect privacy — Don't assume treatment status
9. Inclusive language — Accessible to all
10. Strengths-based — Acknowledge effort and growth

Why it matters:

• Language shapes how users experience the product
• Wrong language can pathologize, shame, or mislead
• Right language validates, supports, and empowers

Enforce through:

• Automated validation (keyword checks)
• Professional review (therapist feedback)
• User testing (diverse feedback)
• Continuous iteration

Get the language right, and you create a genuinely supportive experience.